Search results for: word to vector

Authors: Anudeep Appe, Bhanu Poluparthi, Lakshmi Kasivajjula, Udai Mv, Sobha Bagadi, Punya Modi, Aditya Singh, Hemanth Gunupudi, Spenser Troiano, Jeff Paul, Justin Stovall, Justin Yamamoto

Abstract:

The necessity of data-driven decisions in healthcare strategy formulation is rapidly increasing. A reliable framework which helps identify factors impacting a healthcare provider facility or a hospital (from here on termed as facility) market share is of key importance. This pilot study aims at developing a data-driven machine learning-regression framework which aids strategists in formulating key decisions to improve the facility’s market share which in turn impacts in improving the quality of healthcare services. The US (United States) healthcare business is chosen for the study, and the data spanning 60 key facilities in Washington State and about 3 years of historical data is considered. In the current analysis, market share is termed as the ratio of the facility’s encounters to the total encounters among the group of potential competitor facilities. The current study proposes a two-pronged approach of competitor identification and regression approach to evaluate and predict market share, respectively. Leveraged model agnostic technique, SHAP, to quantify the relative importance of features impacting the market share. Typical techniques in literature to quantify the degree of competitiveness among facilities use an empirical method to calculate a competitive factor to interpret the severity of competition. The proposed method identifies a pool of competitors, develops Directed Acyclic Graphs (DAGs) and feature level word vectors, and evaluates the key connected components at the facility level. This technique is robust since its data-driven, which minimizes the bias from empirical techniques. The DAGs factor in partial correlations at various segregations and key demographics of facilities along with a placeholder to factor in various business rules (for ex. quantifying the patient exchanges, provider references, and sister facilities). Identified are the multiple groups of competitors among facilities. Leveraging the competitors' identified developed and fine-tuned Random Forest Regression model to predict the market share. To identify key drivers of market share at an overall level, permutation feature importance of the attributes was calculated. For relative quantification of features at a facility level, incorporated SHAP (SHapley Additive exPlanations), a model agnostic explainer. This helped to identify and rank the attributes at each facility which impacts the market share. This approach proposes an amalgamation of the two popular and efficient modeling practices, viz., machine learning with graphs and tree-based regression techniques to reduce the bias. With these, we helped to drive strategic business decisions.

Keywords: competition, DAGs, facility, healthcare, machine learning, market share, random forest, SHAP

Procedia PDF Downloads 79

Authors: E. Nakamachi, S. Tanaka, K. Yamamoto, Y. Morita

Abstract:

In this study, we developed a three-dimensional (3D) direct current electric field (DCEF) stimulation bio-reactor for axonal outgrowth enhancement to generate the neural network of the central nervous system (CNS). By using our newly developed 3D DCEF stimulation bio-reactor, we cultured the rat pheochromocytoma cells (PC12) and investigated the effects on the axonal extension enhancement and network generation. Firstly, we designed and fabricated a 3D bio-reactor, which can load DCEF stimulation on PC12 cells embedded in the collagen gel as extracellular environment. The connection between the electrolyte and the medium using salt bridges for DCEF stimulation was introduced to avoid the cell death by the toxicity of metal ion. The distance between the salt bridges was adopted as the design variable to optimize a structure for uniform DCEF stimulation, where the finite element (FE) analyses results were used. Uniform DCEF strength and electric flux vector direction in the PC12 cells embedded in collagen gel were examined through measurements of the fabricated 3D bio-reactor chamber. Measurement results of DCEF strength in the bio-reactor showed a good agreement with FE results. In addition, the perfusion system was attached to maintain pH 7.2 ~ 7.6 of the medium because pH change was caused by DCEF stimulation loading. Secondly, we disseminated PC12 cells in collagen gel and carried out 3D culture. Finally, we measured the morphology of PC12 cell bodies and neurites by the multiphoton excitation fluorescence microscope (MPM). The effectiveness of DCEF stimulation to enhance the axonal outgrowth and the neural network generation was investigated. We confirmed that both an increase of mean axonal length and axogenesis rate of PC12, which have been exposed 5 mV/mm for 6 hours a day for 4 days in the bioreactor. We found following conclusions in our study. 1) Design and fabrication of DCEF stimulation bio-reactor capable of 3D culture nerve cell were completed. A uniform electric field strength of average value of 17 mV/mm within the 1.2% error range was confirmed by using FE analyses, after the structure determination through the optimization process. In addition, we attached a perfusion system capable of suppressing the pH change of the culture solution due to DCEF stimulation loading. 2) Evaluation of DCEF stimulation effects on PC12 cell activity was executed. The 3D culture of PC 12 was carried out adopting the embedding culture method using collagen gel as a scaffold for four days under the condition of 5.0 mV/mm and 10mV/mm. There was a significant effect on the enhancement of axonal extension, as 11.3% increase in an average length, and the increase of axogenesis rate. On the other hand, no effects on the orientation of axon against the DCEF flux direction was observed. Further, the network generation was enhanced to connect longer distance between the target neighbor cells by DCEF stimulation.

Keywords: PC12, DCEF stimulation, 3D bio-reactor, axonal extension, neural network generation

Procedia PDF Downloads 178

Authors: Ulysse Lecomte, Maryline Thenot

Abstract:

For a long time, disabled people lived in isolation within the family environment, with little interaction with the outside world and a high risk of social exclusion. However, in a number of countries, progress has been made thanks to changes in legislation on the social integration of disabled people, a significant change in attitudes, and the development of CSR. But the problem of their social, economic, and professional exclusion persists and has been further exacerbated by the COVID-19 pandemic. This societal phenomenon is sufficiently important to be the subject of management science research. We have therefore focused our work on society's current perception of people with disabilities and their possible integration. Our aim is to find out what levers could be put in place to bring about positive change in the situation. We have chosen to focus on the perception of young people in France, who are the new generation responsible for the future of our society and from whom tomorrow's decisionmakers, future employers, and stakeholders who can influence the living conditions of disabled people will be drawn. Our study sample corresponds to the 18-30 age group, which is the population of young adults likely to have sufficient experience and maturity. The aim of this study is not only to find out how this population currently perceives disability but also to identify the factors influencing this perception and the most effective levers for action to act positively on this phenomenon and thus promote better social integration of people with disabilities in the future. The methodology is based on theoretical and empirical research. The literature review includes a historical and etymological approach to disability, a definition of the different concepts of disability, an approach to disability as a vector of social exclusion, and the role of perception and representations in defining the social image of disability. This literature review is followed by an empirical part carried out by means of a questionnaire administered to 110 young people aged 18 to 30. Analysis of our results suggests that, despite a recent improvement, disabled people are still perceived as vulnerable and socially marginalised. The following factors stand out as having a significant influence (positive or negative) on the perception of disability: the individual's familiarity with the 'world of disability', cultural factors, the degree of 'visibility' of the disability and the empathy level of the disabled person him/herself. Others, on the other hand, such as socio-political and economic factors, have little impact on this perception. In addition, it is possible to classify the various levers of action likely to improve the social perception of disability according to their degree of effectiveness. Our study population prioritised training initiatives for the various players and stakeholders (teachers, students, disabled people themselves, companies, sports clubs, etc.). This was followed by communication, ecommunication and media campaigns in favour of disability. Lastly, the sample was judged as 'less effective' positive discrimination actions such as setting a minimum percentage for the representation of disabled people in various fields (studies, employment, politics ...).

Keywords: disability, perception, social image, young people, influencing factors, levers for action

Procedia PDF Downloads 26

Authors: Chin-Chia Jane

Abstract:

In a transportation network, travel time refers to the transmission time from source node to destination node, whereas reliability refers to the probability of a successful connection from source node to destination node. With an increasing emphasis on quality of service (QoS), both performance indexes are significant in the design and analysis of transportation systems. In this work, we extend the well-known flow network model for transportation networks so that travel time and reliability are integrated into the QoS measurement simultaneously. In the extended model, in addition to the general arc capacities, each intermediate node has a time weight which is the travel time for per unit of commodity going through the node. Meanwhile, arcs and nodes are treated as binary random variables that switch between operation and failure with associated probabilities. For pre-specified travel time limitation and demand requirement, the QoS of a transportation network is the probability that source can successfully transport the demand requirement to destination while the total transmission time is under the travel time limitation. This work is pioneering, since existing literatures that evaluate travel time reliability via a single optimization path, the proposed QoS focuses the performance of the whole network system. To compute the QoS of transportation networks, we first transfer the extended network model into an equivalent min-cost max-flow network model. In the transferred network, each arc has a new travel time weight which takes value 0. Each intermediate node is replaced by two nodes u and v, and an arc directed from u to v. The newly generated nodes u and v are perfect nodes. The new direct arc has three weights: travel time, capacity, and operation probability. Then the universal set of state vectors is recursively decomposed into disjoint subsets of reliable, unreliable, and stochastic vectors until no stochastic vector is left. The decomposition is made possible by applying existing efficient min-cost max-flow algorithm. Because the reliable subsets are disjoint, QoS can be obtained directly by summing the probabilities of these reliable subsets. Computational experiments are conducted on a benchmark network which has 11 nodes and 21 arcs. Five travel time limitations and five demand requirements are set to compute the QoS value. To make a comparison, we test the exhaustive complete enumeration method. Computational results reveal the proposed algorithm is much more efficient than the complete enumeration method. In this work, a transportation network is analyzed by an extended flow network model where each arc has a fixed capacity, each intermediate node has a time weight, and both arcs and nodes are independent binary random variables. The quality of service of the transportation network is an integration of customer demands, travel time, and the probability of connection. We present a decomposition algorithm to compute the QoS efficiently. Computational experiments conducted on a prototype network show that the proposed algorithm is superior to existing complete enumeration methods.

Keywords: quality of service, reliability, transportation network, travel time

Procedia PDF Downloads 211

Authors: Sun-Hee Lee, Amanda Kraley

Abstract:

The “MeToo” movement is a social movement against sexual abuse and harassment. Though the hashtag went viral in 2017 following different cultural flashpoints in different countries, the initial response was quiet in South Korea. This radically changed in January 2018, when a high-ranking senior prosecutor, Seo Ji-hyun, gave a televised interview discussing being sexually assaulted by a colleague. Acknowledging public anger, particularly among women, on the long-existing problems of sexual harassment and abuse, the South Korean media have focused on several high-profile cases. Analyzing the media representation of these cases is a window into the evolving South Korean discourse around “MeToo.” This study presents a linguistic analysis of “MeToo” discourse in South Korea by utilizing a corpus-based approach. The term corpus (pl. corpora) is used to refer to electronic language data, that is, any collection of recorded instances of spoken or written language. A “MeToo” corpus has been collected by extracting newspaper articles containing the keyword “MeToo” from BIGKinds, big data analysis, and service and Nexis Uni, an online academic database search engine, to conduct this language analysis. The corpus analysis explores how Korean media represent accusers and the accused, victims and perpetrators. The extracted data includes 5,885 articles from four broadsheet newspapers (Chosun, JoongAng, Hangyore, and Kyunghyang) and 88 articles from two Korea-based English newspapers (Korea Times and Korea Herald) between January 2017 and November 2020. The information includes basic data analysis with respect to keyword frequency and network analysis and adds refined examinations of select corpus samples through naming strategies, semantic relations, and pragmatic properties. Along with the exponential increase of the number of articles containing the keyword “MeToo” from 104 articles in 2017 to 3,546 articles in 2018, the network and keyword analysis highlights ‘US,’ ‘Harvey Weinstein’, and ‘Hollywood,’ as keywords for 2017, with articles in 2018 highlighting ‘Seo Ji-Hyun, ‘politics,’ ‘President Moon,’ ‘An Ui-Jeong, ‘Lee Yoon-taek’ (the names of perpetrators), and ‘(Korean) society.’ This outcome demonstrates the shift of media focus from international affairs to domestic cases. Another crucial finding is that word ‘defamation’ is widely distributed in the “MeToo” corpus. This relates to the South Korean legal system, in which a person who defames another by publicly alleging information detrimental to their reputation—factual or fabricated—is punishable by law (Article 307 of the Criminal Act of Korea). If the defamation occurs on the internet, it is subject to aggravated punishment under the Act on Promotion of Information and Communications Network Utilization and Information Protection. These laws, in particular, have been used against accusers who have publicly come forward in the wake of “MeToo” in South Korea, adding an extra dimension of risk. This corpus analysis of “MeToo” newspaper articles contributes to the analysis of the media representation of the “MeToo” movement and sheds light on the shifting landscape of gender relations in the public sphere in South Korea.

Keywords: corpus linguistics, MeToo, newspapers, South Korea

Procedia PDF Downloads 206

Authors: Shpetim Rezniqi

Abstract:

Financial Policies in the process of global crisis the current crisis has swept the world with special emphasis, most developed countries, those countries which have most gross -product world and you have a high level of living.Even those who are not experts can describe the consequences of the crisis to see the reality that is seen, but how far will it go this crisis is impossible to predict. Even the biggest experts have conjecture and large divergence, but agree on one thing: - The devastating effects of this crisis will be more severe than ever before and can not be predicted.Long time, the world was dominated economic theory of free market laws. With the belief that the market is the regulator of all economic problems. The market, as river water will flow to find the best and will find the necessary solution best. Therefore much less state market barriers, less state intervention and market itself is an economic self-regulation. Free market economy became the model of global economic development and progress, it transcends national barriers and became the law of the development of the entire world economy. Globalization and global market freedom were principles of development and international cooperation. All international organizations like the World Bank, states powerful economic, development and cooperation principles laid free market economy and the elimination of state intervention. The less state intervention much more freedom of action was this market- leading international principle. We live in an era of financial tragic. Financial markets and banking in particular economies are in a state of thy good, US stock markets fell about 40%, in other words, this time, was one of the darkest moments 5 since 1920. Prior to her rank can only "collapse" of the stock of Wall Street in 1929, technological collapse of 2000, the crisis of 1973 after the Yom Kippur war, while the price of oil quadrupled and famous collapse of 1937 / '38, when Europe was beginning World war II In 2000, even though it seems like the end of the world was the corner, the world economy survived almost intact. Of course, that was small recessions in the United States, Europe, or Japan. Much more difficult the situation was at crisis 30s, or 70s, however, succeeded the world. Regarding the recent financial crisis, it has all the signs to be much sharper and with more consequences. The decline in stock prices is more a byproduct of what is really happening. Financial markets began dance of death with the credit crisis, which came as a result of the large increase in real estate prices and household debt. It is these last two phenomena can be matched very well with the gains of the '20s, a period during which people spent fists as if there was no tomorrow. All is not away from the mouth of the word recession, that fact no longer a sudden and abrupt. But as much as the financial markets melt, the greater is the risk of a problematic economy for years to come. Thus, for example, the banking crisis in Japan proved to be much more severe than initially expected, partly because the assets which were based more loans had, especially the land that falling in value. The price of land in Japan is about 15 years that continues to fall. (ADRI Nurellari-Published in the newspaper "Classifieds"). At this moment, it is still difficult to çmosh to what extent the crisis has affected the economy and what would be the consequences of the crisis. What we know is that many banks will need more time to reduce the award of credit, but banks have this primary function, this means huge loss.

Keywords: globalisation, finance, crisis, recomandation, bank, credits

Procedia PDF Downloads 376

Authors: Qianxi Lv, Junying Liang

Abstract:

Among the various modes of interpreting, simultaneous interpreting (SI) is regarded as a ‘complex’ and ‘extreme condition’ of cognitive tasks while consecutive interpreters (CI) do not have to share processing capacity between tasks. Given that SI exerts great cognitive demand, it makes sense to posit that the output of SI may be more compromised than that of CI in the linguistic features. The bulk of the research has stressed the varying cognitive demand and processes involved in different modes of interpreting; however, related empirical research is sparse. In keeping with our interest in investigating the quantitative linguistic factors discriminating between SI and CI, the current study seeks to examine the potential lexical simplification, syntactic complexity and sequential organization mechanism with a self-made inter-model corpus of transcribed simultaneous and consecutive interpretation, translated speech and original speech texts with a total running word of 321960. The lexical features are extracted in terms of the lexical density, list head coverage, hapax legomena, and type-token ratio, as well as core vocabulary percentage. Dependency distance, an index for syntactic complexity and reflective of processing demand is employed. Frequency motif is a non-grammatically-bound sequential unit and is also used to visualize the local function distribution of interpreting the output. While SI is generally regarded as multitasking with high cognitive load, our findings evidently show that CI may impose heavier or taxing cognitive resource differently and hence yields more lexically and syntactically simplified output. In addition, the sequential features manifest that SI and CI organize the sequences from the source text in different ways into the output, to minimize the cognitive load respectively. We reasoned the results in the framework that cognitive demand is exerted both on maintaining and coordinating component of Working Memory. On the one hand, the information maintained in CI is inherently larger in volume compared to SI. On the other hand, time constraints directly influence the sentence reformulation process. The temporal pressure from the input in SI makes the interpreters only keep a small chunk of information in the focus of attention. Thus, SI interpreters usually produce the output by largely retaining the source structure so as to relieve the information from the working memory immediately after formulated in the target language. Conversely, CI interpreters receive at least a few sentences before reformulation, when they are more self-paced. CI interpreters may thus tend to retain and generate the information in a way to lessen the demand. In other words, interpreters cope with the high demand in the reformulation phase of CI by generating output with densely distributed function words, more content words of higher frequency values and fewer variations, simpler structures and more frequently used language sequences. We consequently propose a revised effort model based on the result for a better illustration of cognitive demand during both interpreting types.

Keywords: cognitive demand, corpus-based, dependency distance, frequency motif, interpreting types, lexical simplification, sequential units distribution, syntactic complexity

Procedia PDF Downloads 160

Authors: Surachart Koyadun

Abstract:

Introduction: Plasmodium knowlesi (P. knowlesi) malaria is a zoonotic disease that is classified as type 5 of human malaria. Commonly found in macaques (Macaca fascicularis) and (Macaca nemestrina), P. knowlesi is capable of resulting in both uncomplicated and severe malaria in humans. Situation of P. knowlesi malaria in Phang-Nga province for the past 3 years from 2020 – 2022 revealed no case report in 2020, however, a total of 14 cases had been reported in 2021 - 2022. This research aimed to 1) study the epidemiology of P. knowlesi, 2) examine the clinical manifestations of P. knowlesi patients, 3) analyze the ecology and entomology of P. knowlesi, and 4) analyze the diagnosis and treatment of P. knowlesi. Method: This research was a retrospective descriptive study/case report. The study was conducted in 14 patients with P. knowlesi malaria between 2021 and 2022 in 4 districts of Phang-Nga Province, Thailand including Thapput, Kapong, Takuapa and Khuraburi. Results: The study subjects of P. knowlesi malaria were all males. Most of them were working age groups as farmers and worked in forest or plantation areas. All had no history of blood transfusions. Most of the patients did not use mosquito nets and had a history of camping in the forest prior to the onset of fever. An analysis of all 14 sources of infection unveiled the area is home to macaques, and that area has detected Anopheles mosquito, which is the carrier of the disease. Majority of them got sick in the dry season of Thailand (December-April). The main symptoms brought to the hospital were fever, chills, headache, body aches. Laboratory findings on the first day of diagnosis were as follows: The white blood cell count was found within the normal range. In the proportion of white blood cells, eosinophils were found to be slightly higher than normal. Slight anemia was found on early examination. The platelet count was found to be below normal in all cases. Severely low platelet count (2,000 cells/mm3) was found in severe cases with multiple complications. No patient was found dead but 85.7% of complications were found, with acute renal failure being the most common. Patients with delayed diagnosis and treatment of malaria (inaccurate diagnosis or late access to the hospital) had the highest severity and complications than those who had seen the doctor since the first 3-4 days of illness or the screening of symptoms and risk history by the malaria clinic staff at vector-borne disease control unit. Conclusion and Recommendation: P. knowlesi malaria is an emerging infectious disease transmitted from animals to humans. There are challenges in epidemiology, entomology, ecology for effective surveillance, prevention and control. Early diagnosis and treatment would reduce complications and prevent death.

Keywords: malaria, plasmodium knowlesi, epidemiology, ecology, entomology, diagnosis, treatment

Procedia PDF Downloads 57

Authors: Roberta Martino, Viviana Ventre

Abstract:

Intertemporal choices are choices under conditions of uncertainty in which the consequences are distributed over time. The Discounted Utility Model is the essential reference for describing the individual in the context of intertemporal choice. The model is based on the idea that the individual selects the alternative with the highest utility, which is calculated by multiplying the cardinal utility of the outcome, as if the reception were instantaneous, by the discount function that determines a decrease in the utility value according to how the actual reception of the outcome is far away from the moment the choice is made. Initially, the discount function was assumed to have an exponential trend, whose decrease over time is constant, in line with a profile of a rational investor described by classical economics. Instead, empirical evidence called for the formulation of alternative, hyperbolic models that better represented the actual actions of the investor. Attitudes that do not comply with the principles of classical rationality are termed anomalous, i.e., difficult to rationalize and describe through normative models. The development of behavioral finance, which describes investor behavior through cognitive psychology, has shown that deviations from rationality are due to the limited rationality condition of human beings. What this means is that when a choice is made in a very difficult and information-rich environment, the brain does a compromise job between the cognitive effort required and the selection of an alternative. Moreover, the evaluation and selection phase of the alternative, the collection and processing of information, are dynamics conditioned by systematic distortions of the decision-making process that are the behavioral biases involving the individual's emotional and cognitive system. In this paper we present an original decomposition of the discount function to investigate the psychological principles of hyperbolic discounting. It is possible to decompose the curve into two components: the first component is responsible for the smaller decrease in the outcome as time increases and is related to the individual's impatience; the second component relates to the change in the direction of the tangent vector to the curve and indicates how much the individual perceives the indeterminacy of the future indicating his or her aversion to uncertainty. This decomposition allows interesting conclusions to be drawn with respect to the concept of impatience and the emotional drives involved in decision-making. The contribution that neuroscience can make to decision theory and inter-temporal choice theory is vast as it would allow the description of the decision-making process as the relationship between the individual's emotional and cognitive factors. Neurofinance is a discipline that uses a multidisciplinary approach to investigate how the brain influences decision-making. Indeed, considering that the decision-making process is linked to the activity of the prefrontal cortex and amygdala, neurofinance can help determine the extent to which abnormal attitudes respect the principles of rationality.

Keywords: impatience, intertemporal choice, neurofinance, rationality, uncertainty

Procedia PDF Downloads 115

Authors: B. R. Campomanes-Alvarez, P. Quiros, B. Fernandez

Abstract:

Automatic Speech Recognition (ASR) is a machine-based process of decoding and transcribing oral speech. A typical ASR system receives acoustic input from a speaker or an audio file, analyzes it using algorithms, and produces an output in the form of a text. Some speech recognition systems use Hidden Markov Models (HMMs) to deal with the temporal variability of speech and Gaussian Mixture Models (GMMs) to determine how well each state of each HMM fits a short window of frames of coefficients that represents the acoustic input. Another way to evaluate the fit is to use a feed-forward neural network that takes several frames of coefficients as input and produces posterior probabilities over HMM states as output. Deep neural networks (DNNs) that have many hidden layers and are trained using new methods have been shown to outperform GMMs on a variety of speech recognition systems. Acoustic models for state-of-the-art ASR systems are usually training on massive amounts of data. However, audio files with their corresponding transcriptions can be difficult to obtain, especially in the Spanish language. Hence, in the case of these low-resource scenarios, building an ASR model is considered as a complex task due to the lack of labeled data, resulting in an under-trained system. Semi-supervised learning approaches arise as necessary tasks given the high cost of transcribing audio data. The main goal of this proposal is to develop a procedure based on acoustic semi-supervised learning for Spanish ASR systems by using DNNs. This semi-supervised learning approach consists of: (a) Training a seed ASR model with a DNN using a set of audios and their respective transcriptions. A DNN with a one-hidden-layer network was initialized; increasing the number of hidden layers in training, to a five. A refinement, which consisted of the weight matrix plus bias term and a Stochastic Gradient Descent (SGD) training were also performed. The objective function was the cross-entropy criterion. (b) Decoding/testing a set of unlabeled data with the obtained seed model. (c) Selecting a suitable subset of the validated data to retrain the seed model, thereby improving its performance on the target test set. To choose the most precise transcriptions, three confidence scores or metrics, regarding the lattice concept (based on the graph cost, the acoustic cost and a combination of both), was performed as selection technique. The performance of the ASR system will be calculated by means of the Word Error Rate (WER). The test dataset was renewed in order to extract the new transcriptions added to the training dataset. Some experiments were carried out in order to select the best ASR results. A comparison between a GMM-based model without retraining and the DNN proposed system was also made under the same conditions. Results showed that the semi-supervised ASR-model based on DNNs outperformed the GMM-model, in terms of WER, in all tested cases. The best result obtained an improvement of 6% relative WER. Hence, these promising results suggest that the proposed technique could be suitable for building ASR models in low-resource environments.

Keywords: automatic speech recognition, deep neural networks, machine learning, semi-supervised learning

Procedia PDF Downloads 327

Authors: Yiea Funk Te, Daniel Mueller, Irena Pletikosa Cvijikj

Abstract:

Small and medium enterprises (SMEs) play an important role in the economy of many countries. When the overall world economy is considered, SMEs represent 95% of all businesses in the world, accounting for 66% of the total employment. Existing studies show that the current business environment is characterized as highly turbulent and strongly influenced by modern information and communication technologies, thus forcing SMEs to experience more severe challenges in maintaining their existence and expanding their business. To support SMEs at improving their competitiveness, researchers recently turned their focus on applying data mining techniques to build risk and growth prediction models. However, data used to assess risk and growth indicators is primarily obtained via questionnaires, which is very laborious and time-consuming, or is provided by financial institutes, thus highly sensitive to privacy issues. Recently, web mining (WM) has emerged as a new approach towards obtaining valuable insights in the business world. WM enables automatic and large scale collection and analysis of potentially valuable data from various online platforms, including companies’ websites. While WM methods have been frequently studied to anticipate growth of sales volume for e-commerce platforms, their application for assessment of SME risk and growth indicators is still scarce. Considering that a vast proportion of SMEs own a website, WM bears a great potential in revealing valuable information hidden in SME websites, which can further be used to understand SME risk and growth indicators, as well as to enhance current SME risk and growth prediction models. This study aims at developing an automated system to collect business-relevant data from the Web and predict future growth trends of SMEs by means of WM and data mining techniques. The envisioned system should serve as an 'early recognition system' for future growth opportunities. In an initial step, we examine how structured and semi-structured Web data in governmental or SME websites can be used to explain the success of SMEs. WM methods are applied to extract Web data in a form of additional input features for the growth prediction model. The data on SMEs provided by a large Swiss insurance company is used as ground truth data (i.e. growth-labeled data) to train the growth prediction model. Different machine learning classification algorithms such as the Support Vector Machine, Random Forest and Artificial Neural Network are applied and compared, with the goal to optimize the prediction performance. The results are compared to those from previous studies, in order to assess the contribution of growth indicators retrieved from the Web for increasing the predictive power of the model.

Keywords: data mining, SME growth, success factors, web mining

Procedia PDF Downloads 249

Authors: Burak Oksuzoglu, Savas Yildirim, Ferhat Kutlu

Abstract:

Opinion mining (OM) is one of the natural language processing (NLP) problems to determine the polarity of opinions, mostly represented on a positive-neutral-negative axis. The data for OM is usually collected from various social media platforms. In an era where social media has considerable control over companies’ futures, it’s worth understanding social media and taking actions accordingly. OM comes to the fore here as the scale of the discussion about companies increases, and it becomes unfeasible to gauge opinion on individual levels. Thus, the companies opt to automize this process by applying machine learning (ML) approaches to their data. For the last two decades, OM or sentiment analysis (SA) has been mainly performed by applying ML classification algorithms such as support vector machines (SVM) and Naïve Bayes to a bag of n-gram representations of textual data. With the advent of deep learning and its apparent success in NLP, traditional methods have become obsolete. Transfer learning paradigm that has been commonly used in computer vision (CV) problems started to shape NLP approaches and language models (LM) lately. This gave a sudden rise to the usage of the pretrained language model (PTM), which contains language representations that are obtained by training it on the large datasets using self-supervised learning objectives. The PTMs are further fine-tuned by a specialized downstream task dataset to produce efficient models for various NLP tasks such as OM, NER (Named-Entity Recognition), Question Answering (QA), and so forth. In this study, the traditional and modern NLP approaches have been evaluated for OM by using a sizable corpus belonging to a large private company containing about 76,000 comments in Turkish: SVM with a bag of n-grams, and two chosen pre-trained models, multilingual universal sentence encoder (MUSE) and bidirectional encoder representations from transformers (BERT). The MUSE model is a multilingual model that supports 16 languages, including Turkish, and it is based on convolutional neural networks. The BERT is a monolingual model in our case and transformers-based neural networks. It uses a masked language model and next sentence prediction tasks that allow the bidirectional training of the transformers. During the training phase of the architecture, pre-processing operations such as morphological parsing, stemming, and spelling correction was not used since the experiments showed that their contribution to the model performance was found insignificant even though Turkish is a highly agglutinative and inflective language. The results show that usage of deep learning methods with pre-trained models and fine-tuning achieve about 11% improvement over SVM for OM. The BERT model achieved around 94% prediction accuracy while the MUSE model achieved around 88% and SVM did around 83%. The MUSE multilingual model shows better results than SVM, but it still performs worse than the monolingual BERT model.

Keywords: BERT, MUSE, opinion mining, pretrained language model, SVM, Turkish

Procedia PDF Downloads 128

Authors: Nutcha Taneepanichskul, Helen C. Hailes, Mark Miodownik

Abstract:

Plastic waste has emerged as a critical global environmental challenge, primarily driven by the prevalent use of conventional plastics derived from petrochemical refining and manufacturing processes in modern packaging. While these plastics serve vital functions, their persistence in the environment post-disposal poses significant threats to ecosystems. Addressing this issue necessitates approaches, one of which involves the development of biodegradable plastics designed to degrade under controlled conditions, such as industrial composting facilities. It is imperative to note that compostable plastics are engineered for degradation within specific environments and are not suited for uncontrolled settings, including natural landscapes and aquatic ecosystems. The full benefits of compostable packaging are realized when subjected to industrial composting, preventing environmental contamination and waste stream pollution. Therefore, effective sorting technologies are essential to enhance composting rates for these materials and diminish the risk of contaminating recycling streams. In this study, it leverage hyperspectral imaging technology (HSI) coupled with advanced machine learning algorithms to accurately identify various types of plastics, encompassing conventional variants like Polyethylene terephthalate (PET), Polypropylene (PP), Low density polyethylene (LDPE), High density polyethylene (HDPE) and biodegradable alternatives such as Polybutylene adipate terephthalate (PBAT), Polylactic acid (PLA), and Polyhydroxyalkanoates (PHA). The dataset is partitioned into three subsets: a training dataset comprising uncontaminated conventional and biodegradable plastics, a validation dataset encompassing contaminated plastics of both types, and a testing dataset featuring real-world packaging items in both pristine and contaminated states. Five distinct machine learning algorithms, namely Partial Least Squares Discriminant Analysis (PLS-DA), Support Vector Machine (SVM), Convolutional Neural Network (CNN), Logistic Regression, and Decision Tree Algorithm, were developed and evaluated for their classification performance. Remarkably, the Logistic Regression and CNN model exhibited the most promising outcomes, achieving a perfect accuracy rate of 100% for the training and validation datasets. Notably, the testing dataset yielded an accuracy exceeding 80%. The successful implementation of this sorting technology within recycling and composting facilities holds the potential to significantly elevate recycling and composting rates. As a result, the envisioned circular economy for plastics can be established, thereby offering a viable solution to mitigate plastic pollution.

Keywords: biodegradable plastics, sorting technology, hyperspectral imaging technology, machine learning algorithms

Procedia PDF Downloads 63

Authors: Yakup Ulusu, Faruk Ozel, Numan Eczacioglu, Abdurrahman Ozen, Sabriye Acikgoz

Abstract:

GFP discovered in the mid-1970s, has been used as a marker after replicated genetic study by scientists. In biotechnology, cell, molecular biology, the GFP gene is frequently used as a reporter of expression. In modified forms, it has been used to make biosensors. Many animals have been created that express GFP as an evidence that a gene can be expressed throughout a given organism. Proteins labeled with GFP identified locations are determined. And so, cell connections can be monitored, gene expression can be reported, protein-protein interactions can be observed and signals that create events can be detected. Additionally, monitoring GFP is noninvasive; it can be detected by under UV-light because of simply generating fluorescence. Moreover, GFP is a relatively small and inert molecule, that does not seem to treat any biological processes of interest. The synthesis of GFP has some steps like, to construct the plasmid system, transformation in E. coli, production and purification of protein. GFP carrying plasmid vector pBAD–GFPuv was digested using two different restriction endonuclease enzymes (NheI and Eco RI) and DNA fragment of GFP was gel purified before cloning. The GFP-encoding DNA fragment was ligated into pET28a plasmid using NheI and Eco RI restriction sites. The final plasmid was named pETGFP and DNA sequencing of this plasmid indicated that the hexa histidine-tagged GFP was correctly inserted. Histidine-tagged GFP was expressed in an Escherichia coli BL21 DE3 (pLysE) strain. The strain was transformed with pETGFP plasmid and grown on LuiraBertoni (LB) plates with kanamycin and chloramphenicol selection. E. coli cells were grown up to an optical density (OD 600) of 0.8 and induced by the addition of a final concentration of 1mM isopropyl-thiogalactopyranoside (IPTG) and then grown for additional 4 h. The amino-terminal hexa-histidine-tag facilitated purification of the GFP by using a His Bind affinity chromatography resin (Novagen). Purity of GFP protein was analyzed by a 12 % sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). The concentration of protein was determined by UV absorption at 280 nm (Varian Cary 50 Scan UV/VIS spectrophotometer). Synthesis of GFP-Polymer composite nanofibers was produced by using GFP solution (10mg/mL) and polymer precursor Polyvinylpyrrolidone, (PVP, Mw=1300000) as starting materials and template, respectively. For the fabrication of nanofibers with the different fiber diameter; a sol–gel solution comprising of 0.40, 0.60 and 0.80 g PVP (depending upon the desired fiber diameter) and 100 mg GFP in 10 mL water: ethanol (3:2) mixtures were prepared and then the solution was covered on collecting plate via electro spinning at 10 kV with a feed-rate of 0.25 mL h-1 using Spellman electro spinning system. Results show that GFP-based nano-fiber can be used plenty of biomedical applications such as bio-imaging, bio-mechanic, bio-material and tissue engineering.

Keywords: biomaterial, GFP, nano-fibers, protein expression

Procedia PDF Downloads 308

Authors: Rama Munajat

Abstract:

Discourse information levels may be viewed from either a background-foreground distinction or a multi-level perspective, and cross-linguistic studies on this area suggest that each information level is marked by a specific linguistic device. In this sense, Sundanese, spoken in Indonesia’s West Javanese Province, further differentiates the background and foreground information into ordinary and significant types. This paper will report an ongoing shift from particle-to-active voice marking in the way Sundanese signals foregrounding events. The shift relates to decades of contact with Bahasa Indonesia (Indonesia’s official language) and linguistic compatibility between the two surface marking strategies. Representative data analyzed include three groups of short stories in both Sundanese and Bahasa Indonesia (Indonesian) published in three periods: before 1945, 1965-2006, and 2016-2019. In the first group of Sundanese data, forward-moving events dominantly appear in particle KA (Kecap Anteuran, word-accompanying) constructions, where the KA represents different particles that co-occur with a special group of verbs. The second group, however, shows that the foregrounded events are more frequently described in active-voice forms with a subject-predicate (SP) order. Subsequently, the third offers stronger evidence for the use of the SP structure. As for the Indonesian data, the foregrounding events in the first group occur in verb-initial and passive-voice constructions, while in the second and third, the events more frequently appear in active-voice structures (subject-predicate sequence). The marking shift above suggests a structural influence from Indonesian, stemmed from generational differences among authors of the Sundanese short stories, particularly related to their education and language backgrounds. The first group of short stories – published before 1945 or before Indonesia's independence from Dutch – were written by native speakers of Sundanese who spoke Indonesian as a foreign language and went through the Dutch education system. The second group of authors, on the other hand, represents a generation of Sundanese native speakers who spoke Indonesian as a second language. Finally, the third group consists of authors who are bilingual speakers of both Sundanese and Indonesian. The data suggest that the last two groups of authors completed the Indonesian education system. With these, the use of subject-predicate sequences to denote foregrounding events began to appear more frequently in the second group and then became more dominant in those of the third. The coded data also signify that cohesion, coherence, and pragmatic purposes in Particle KA constructions are intact in their respective active-voice structure counterparts. For instance, the foregrounding events in Particle KA constructions occur in Sentence-initial KA and Pre-verbal KA forms, whereas those in the active-voice are described in Subject-Predicate (SP) and Zero-Subject active-voice patterns. Cross-language data further demonstrate that the Sentence-initial KA and the SP active-voice structures each contain an overt noun phrase (NP) co-referential with one of the entities introduced in a preceding context. Similarly, the pre-verbal KA and Zero-Subject active-voice patterns have a deleted noun phrase unambiguously referable to the only one entity previously mentioned. The presence and absence of an NP inform a pragmatic strategy to place prominence on topic/given and comment/new information, respectively.

Keywords: discourse analysis, foregrounding marking, pragmatics, language contact

Procedia PDF Downloads 122

Authors: Muhammad Alhada Fuadilah Habib, Ayla Karina Budita, Cut Rizka Al Usrah, Mukhammad Fatkhullah, Kanita Khoirun Nisa, Siti Muslihatul Mukaromah

Abstract:

In this contemporary era, the term of development often emphasised merely on the economic growth aspect. Development of a program often considered as superior by the government, in fact, it often raises various problems. The problems occur because the development policies determined by the government tend to favor private entrepreneurs and impose on the oppression toward the community. The development promised to prosper the community's life, turn out in fact of harming the community, threatening the survival of the community and damaging the ecosystem of nature where the community hangs their life to it. Nowadays many natural resources should be used for the community’s life prosperity. However, the prosperity is conquered by the private entrepreneurs that are regulated through the free market mechanism and wrapped in democratization. This condition actually is a form of neoliberalism that builds new administration order system which is far from the meaning of the word democracy. The government should play more role in protecting community's life and prosperity, but in fact, the government sides with the private entrepreneurs for the sake of the economic benefits regardless of other aspects of the community’s life. This unjustified condition presents a wide range of social movements from the community in response to the neoliberalis policy that actually eliminates the doctrine of community sovereignty. Social movements performed by Yogyakarta south coastal area community, as the focus of the discussion in this paper, is one of the community’s response toward the government policies related to the construction of the ferruginous sand quarry which is tend to favor on private entrepreneurs and highly prejudicing or even threatening the survival of Yogyakarta south coastal area community. The data collection in this study uses qualitative research methods with in-depth interview data collection techniques and purposive informant determination techniques. This method was chosen in order to obtain the insightful data and detailed information to uncover the injustice policies committed by the government-private entrepreneurs toward Yogyakarta south coastal area community. The brief results of this study show that the conflicts between the community and government-private entrepreneurs occurred because of the differences of interests and paradigm of natural resource management. The resistance movements done by the community to fight back the government-private entrepreneurs was conducted by forming an organization called Paguyupan Petani Lahan Pantai Kulon Progo (PPLP-KP). This organization do the resistances through two ways; firstly, quiet action done through various actions such as; refusing against the socialization, performing discussion to deliberate their argument with the government-private entrepreneurs, complaining the problems to the central government, creating banners or billboards which contain the writing of rejection, performing pray rituals to invoke the justice from the God, as well as instill the resistance ideology to their young generation. Secondly, the rough action also is done through various actions such as; doing roadblocks, conducting rallies, as well as doing clash with the government apparatus. In case the resistances done by the community are seen from the pattern. Actually, the resistances are reaction toward the aggression carried out by the government-private entrepreneurs.

Keywords: community resistance, conflict, ferruginous sand quarry construction, social movement

Procedia PDF Downloads 272

Authors: Getaneh Berie Tarekegn, Li-Chia Tai

Abstract:

With the rapid advancement of artificial intelligence, low-power built-in sensors on Internet of Things devices, and communication technologies, location-aware services have become increasingly popular and have permeated every aspect of people’s lives. Global navigation satellite systems (GNSSs) are the default method of providing continuous positioning services for ground and aerial vehicles, as well as consumer devices (smartphones, watches, notepads, etc.). However, the environment affects satellite positioning systems, particularly indoors, in dense urban and suburban cities enclosed by skyscrapers, or when deep shadows obscure satellite signals. This is because (1) indoor environments are more complicated due to the presence of many objects surrounding them; (2) reflection within the building is highly dependent on the surrounding environment, including the positions of objects and human activity; and (3) satellite signals cannot be reached in an indoor environment, and GNSS doesn't have enough power to penetrate building walls. GPS is also highly power-hungry, which poses a severe challenge for battery-powered IoT devices. Due to these challenges, IoT applications are limited. Consequently, precise, seamless, and ubiquitous Positioning, Navigation and Timing (PNT) systems are crucial for many artificial intelligence Internet of Things (AI-IoT) applications in the era of smart cities. Their applications include traffic monitoring, emergency alarms, environmental monitoring, location-based advertising, intelligent transportation, and smart health care. This paper proposes a generative AI-based positioning scheme for large-scale wireless settings using fingerprinting techniques. In this article, we presented a semi-supervised deep convolutional generative adversarial network (S-DCGAN)-based radio map construction method for real-time device localization. We also employed a reliable signal fingerprint feature extraction method with t-distributed stochastic neighbor embedding (t-SNE), which extracts dominant features while eliminating noise from hybrid WLAN and long-term evolution (LTE) fingerprints. The proposed scheme reduced the workload of site surveying required to build the fingerprint database by up to 78.5% and significantly improved positioning accuracy. The results show that the average positioning error of GAILoc is less than 0.39 m, and more than 90% of the errors are less than 0.82 m. According to numerical results, SRCLoc improves positioning performance and reduces radio map construction costs significantly compared to traditional methods.

Keywords: location-aware services, feature extraction technique, generative adversarial network, long short-term memory, support vector machine

Procedia PDF Downloads 30

Authors: Yohanna Panshak

Abstract:

The impact of macroeconomic instability on economic growth and prosperity has been at forefront in many discourses among researchers and policy makers and has generated a lot of controversies over the years. This has generated series of research efforts towards understanding the remote causes of this phenomenon; its nature, determinants and how it can be targeted and mitigated. While others have opined that the root cause of macroeconomic flux in Nigeria is attributed to Oil-Price volatility, others viewed the issue as resulting from some constellation of structural constraints both within and outside the shores of the country. Research works of scholars such as [Akpan (2009), Aliyu (2009), Olomola (2006), etc] argue that oil volatility can determine economic growth or has the potential of doing so. On the contrary, [Darby (1982), Cerralo (2005) etc] share the opinion that it can slow down growth. The earlier argument rest on the understanding that for a net balance of oil exporting economies, price upbeat directly increases real national income through higher export earnings, whereas, the latter allude to the case of net-oil importing countries (which experience price rises, increased input costs, reduced non-oil demand, low investment, fall in tax revenues and ultimately an increase in budget deficit which will further reduce welfare level). Therefore, assessing the precise impact of oil price volatility on virtually any economy is a function of whether it is an oil-exporting or importing nation. Research on oil price volatility and its outcome on the growth of the Nigerian economy are evolving and in a march towards resolving Nigeria’s macroeconomic instability as long as oil revenue still remain the mainstay and driver of socio-economic engineering. Recently, a major importer of Nigeria’s oil- United States made a historic breakthrough in more efficient source of energy for her economy with the capacity of serving significant part of the world. This undoubtedly suggests a threat to the exchange earnings of the country. The need to understand fluctuation in its major export commodity is critical. This paper leans on the Renaissance growth theory with greater focus on theoretical work of Lee (1998); a leading proponent of this school who makes a clear cut of difference between oil price changes and oil price volatility. Based on the above background, the research seeks to empirically examine the impact oil-price volatility on government expenditure using quarterly time series data spanning 1986:1 to 2014:4. Vector Auto Regression (VAR) econometric approach shall be used. The structural properties of the model shall be tested using Augmented Dickey-Fuller and Phillips-Perron. Relevant diagnostics tests of heteroscedasticity, serial correlation and normality shall also be carried out. Policy recommendation shall be offered on the empirical findings and believes it assist policy makers not only in Nigeria but the world-over.

Keywords: oil-price, volatility, prosperity, budget, expenditure

Procedia PDF Downloads 259

Authors: Nahed Mohamed Ismail, Bayomy Mostafa, Ahmed Abdel Kader, Ahmed Mohamed Azzam

Abstract:

Effect of lining watercourses on the distribution and abundance of fresh water snails at two Egyptian governorates, Baheria (new reclaimed area) and Giza was studied. Seasonal survey in lined and unlined sites during two successive years was carried out. Samples of snails and water were collected from each examined site and the ecological conditions were recorded. The collected snails from each site were placed in plastic aquaria and transferred to the laboratory, where they were sorted out, identified, counted and examined for natural infection. The size frequency distribution was calculated for each snail species. Results revealed that snails were represented in all examined watercourses (lined and unlined) at the two tested habitats by 14 species. (Biomphalaria alexandrina, B. glabrata, Bulinus truncatus, Physa acuta. Helisoma duryi, Lymnaea natalensis, Planorbis planorbis, Cleopatra bulimoids, Lanistes carinatus, Bellamya unicolor, Melanoides tuberculata, Theodoxus nilotica, Succinia cleopatra and Gabbiella senaarensis). During spring, the percentage of live (45%) and dead (55%) snail species was extremely highly significant lower (p>0.001) in lined water bodies compared to the unlined ones (93.5% and 6.5%, respectively) in the examined sites at Baheria. At Giza, the percentage values of live snail species from all lined watercourses (82.6% and 60.2%, during winter and spring, respectively) was significantly lower (p>0.05 & p>0.01) than those in unlined ones (91.1% and 79%, respectively). Size frequency distribution of snails collected from the lined and unlined water bodies at Baheria and Giza governorates during all seasons revealed that during survey, snail populations were stable and the recruitment of young to adult was continuing for some species, where the recruits were observed with adults. However, there was no sign of small snails occurrence in case of B. glabrata and B. alexandrina during autumn, winter and spring and disappear during summer at Giza. Meanwhile they completely absent during all seasons at Baheria Governorate. Chemical analysis of some heavy metals of water samples collected from lined and unlined sites from Baheria and Giza governorates during autumn, winter and spring were approximately as the same in both lined and unlined water bodies. However, Zn and Fe were higher in lined sites (0.78±0.37and 17.4 ± 4.3, respectively) than that of unlined ones (0.4±0.1 and 10.95 ± 1.93, respectively) and Cu was absent in both lined and unlined sites during summer at Baheria governorate. At Giza, Cu and Pb were absent and Fe were higher in lined sites (4.7± 4.2) than that of unlined ones (2.5 ± 1.4) during summer. Statistical analysis showed that no significant difference in all physico-chemical parameters of water in lined and unlined water bodies at the two tested habitats during all seasons. However, it was found that the water conductivity and TDS showed a lower mean values in lined sites than those of unlined ones. Thus, the present obtained data support the concept of utilizing environmental modification such as lining of water courses to help in minimizing the population density of certain vector snails and consequently reduce the transmission of snails born diseases.

Keywords: lining, fresh water, snails, watercourses

Procedia PDF Downloads 240

Authors: Getaneh Berie Tarekegn, Li-Chia Tai

Abstract:

With the rapid advancement of artificial intelligence, low-power built-in sensors on Internet of Things devices, and communication technologies, location-aware services have become increasingly popular and have permeated every aspect of people’s lives. Global navigation satellite systems (GNSSs) are the default method of providing continuous positioning services for ground and aerial vehicles, as well as consumer devices (smartphones, watches, notepads, etc.). However, the environment affects satellite positioning systems, particularly indoors, in dense urban and suburban cities enclosed by skyscrapers, or when deep shadows obscure satellite signals. This is because (1) indoor environments are more complicated due to the presence of many objects surrounding them; (2) reflection within the building is highly dependent on the surrounding environment, including the positions of objects and human activity; and (3) satellite signals cannot be reached in an indoor environment, and GNSS doesn't have enough power to penetrate building walls. GPS is also highly power-hungry, which poses a severe challenge for battery-powered IoT devices. Due to these challenges, IoT applications are limited. Consequently, precise, seamless, and ubiquitous Positioning, Navigation and Timing (PNT) systems are crucial for many artificial intelligence Internet of Things (AI-IoT) applications in the era of smart cities. Their applications include traffic monitoring, emergency alarming, environmental monitoring, location-based advertising, intelligent transportation, and smart health care. This paper proposes a generative AI-based positioning scheme for large-scale wireless settings using fingerprinting techniques. In this article, we presented a novel semi-supervised deep convolutional generative adversarial network (S-DCGAN)-based radio map construction method for real-time device localization. We also employed a reliable signal fingerprint feature extraction method with t-distributed stochastic neighbor embedding (t-SNE), which extracts dominant features while eliminating noise from hybrid WLAN and long-term evolution (LTE) fingerprints. The proposed scheme reduced the workload of site surveying required to build the fingerprint database by up to 78.5% and significantly improved positioning accuracy. The results show that the average positioning error of GAILoc is less than 0.39 m, and more than 90% of the errors are less than 0.82 m. According to numerical results, SRCLoc improves positioning performance and reduces radio map construction costs significantly compared to traditional methods.

Keywords: location-aware services, feature extraction technique, generative adversarial network, long short-term memory, support vector machine

Procedia PDF Downloads 31

Authors: Patrizia Frontera, Anastasia Macario, Sebastiano Candamano, Fortunato Crea, Pierluigi Antonucci

Abstract:

The increasing of the world energy demand makes today biomass an attractive energy source, based on the minimizing of CO2 emission and on the global warming reduction purposes. Recently, COP-21, the international meeting on global climate change, defined the roadmap for sustainable worldwide development, based on low-carbon containing fuel. Hydrogen is an energy vector able to substitute the conventional fuels from petroleum. Ethanol for hydrogen production represents a valid alternative to the fossil sources due to its low toxicity, low production costs, high biodegradability, high H2 content and renewability. Ethanol conversion to generate hydrogen by a combination of partial oxidation and steam reforming reactions is generally called auto-thermal reforming (ATR). The ATR process is advantageous due to the low energy requirements and to the reduced carbonaceous deposits formation. Catalyst plays a pivotal role in the ATR process, especially towards the process selectivity and the carbonaceous deposits formation. Bimetallic or trimetallic catalysts, as well as catalysts with doped-promoters supports, may exhibit high activity, selectivity and deactivation resistance with respect to the corresponding monometallic ones. In this work, NiMoCo/GDC, NiMoCu/GDC and NiMoRe/GDC (where GDC is Gadolinia Doped Ceria support and the metal composition is 60:30:10 for all catalyst) have been prepared by impregnation method. The support, Gadolinia 0.2 Doped Ceria 0.8, was impregnated by metal precursors solubilized in aqueous ethanol solution (50%) at room temperature for 6 hours. After this, the catalysts were dried at 100°C for 8 hours and, subsequently, calcined at 600°C in order to have the metal oxides. Finally, active catalysts were obtained by reduction procedure (H2 atmosphere at 500°C for 6 hours). All sample were characterized by different analytical techniques (XRD, SEM-EDX, XPS, CHNS, H2-TPR and Raman Spectorscopy). Catalytic experiments (auto-thermal reforming of ethanol) were carried out in the temperature range 500-800°C under atmospheric pressure, using a continuous fixed-bed microreactor. Effluent gases from the reactor were analyzed by two Varian CP4900 chromarographs with a TCD detector. The analytical investigation focused on the preventing of the coke deposition, the metals sintering effect and the sulfur poisoning. Hydrogen productivity, ethanol conversion and products distribution were measured and analyzed. At 600°C, all tri-metallic catalysts show the best performance: H2 + CO reaching almost the 77 vol.% in the final gases. While NiMoCo/GDC catalyst shows the best selectivity to hydrogen whit respect to the other tri-metallic catalysts (41 vol.% at 600°C). On the other hand, NiMoCu/GDC and NiMoRe/GDC demonstrated high sulfur poisoning resistance (up to 200 cc/min) with respect to the NiMoCo/GDC catalyst. The correlation among catalytic results and surface properties of the catalysts will be discussed.

Keywords: catalysts, ceria, ethanol, gadolinia, hydrogen, Nickel

Procedia PDF Downloads 144

Authors: Qi Liu, Jiqin Yao, Xiaohu Zhou, Bo Zheng

Abstract:

The first artificial cell was produced by Thomas Chang in the 1950s when he was trying to make a mimic of red blood cells. Since then, many different types of artificial cells have been constructed from one of the two approaches: a so-called bottom-up approach, which aims to create a cell from scratch, and a top-down approach, in which genes are sequentially knocked out from organisms until only the minimal genome required for sustaining life remains. In this project, bottom-up approach was used to build a new cell-free expression system which mimics artificial cell that capable of protein expression and communicate with each other. The artificial cells constructed from the bottom-up approach are usually lipid vesicles, polymersomes, hydrogels or aqueous droplets containing the nucleic acids and transcription-translation machinery. However, lipid vesicles based artificial cells capable of communication present several issues in the cell communication research: (1) The lipid vesicles normally lose the important functions such as protein expression within a few hours. (2) The lipid membrane allows the permeation of only small molecules and limits the types of molecules that can be sensed and released to the surrounding environment for chemical communication; (3) The lipid vesicles are prone to rupture due to the imbalance of the osmotic pressure. To address these issues, the hydrogel-based artificial cells were constructed in this work. To construct the artificial cell, polyacrylamide hydrogel was functionalized with Acrylate PEG Succinimidyl Carboxymethyl Ester (ACLT-PEG2000-SCM) moiety on the polymer backbone. The proteinaceous factors can then be immobilized on the polymer backbone by the reaction between primary amines of proteins and N-hydroxysuccinimide esters (NHS esters) of ACLT-PEG2000-SCM, the plasmid template and ribosome were encapsulated inside the hydrogel particles. Because the artificial cell could continuously express protein with the supply of nutrients and energy, the artificial cell-artificial cell communication and artificial cell-natural cell communication could be achieved by combining the artificial cell vector with designed plasmids. The plasmids were designed referring to the quorum sensing (QS) system of bacteria, which largely relied on cognate acyl-homoserine lactone (AHL) / transcription pairs. In one communication pair, “sender” is the artificial cell or natural cell that can produce AHL signal molecule by synthesizing the corresponding signal synthase that catalyzed the conversion of S-adenosyl-L-methionine (SAM) into AHL, while the “receiver” is the artificial cell or natural cell that can sense the quorum sensing signaling molecule form “sender” and in turn express the gene of interest. In the experiment, GFP was first immobilized inside the hydrogel particle to prove that the functionalized hydrogel particles could be used for protein binding. After that, the successful communication between artificial cell-artificial cell and artificial cell-natural cell was demonstrated, the successful signal between artificial cell-artificial cell or artificial cell-natural cell could be observed by recording the fluorescence signal increase. The hydrogel-based artificial cell designed in this work can help to study the complex communication system in bacteria, it can also be further developed for therapeutic applications.

Keywords: artificial cell, cell-free system, gene circuit, synthetic biology

Procedia PDF Downloads 134

Authors: Siham Boulaknadel

Abstract:

Every culture has its share of a rich history of storytelling in oral, visual, and textual form. The Amazigh language, as many languages, has its own which has entertained and informed across centuries and cultures, and its instructional potential continues to serve teachers. According to many researchers, listening to stories draws attention to the sounds of language and helps children develop sensitivity to the way language works. Stories including repetitive phrases, unique words, and enticing description encourage students to join in actively to repeat, chant, sing, or even retell the story. This kind of practice is important to language learners’ oral language development, which is believed to correlate completely with student’s academic success. Today, with the advent of multimedia, digital storytelling for instance can be a practical and powerful learning tool. It has the potential in transforming traditional learning into a world of unlimited imaginary environment. This paper reports on a research project on development of multimedia Storytelling Website using traditional Amazigh oral narratives called “tell me a story”. It is a didactic tool created for the learning of good moral values in an interactive multimedia environment combining on-screen text, graphics and audio in an enticing environment and enabling the positive values of stories to be projected. This Website developed in this study is based on various pedagogical approaches and learning theories deemed suitable for children age 8 to 9 year-old. The design and development of Website was based on a well-researched conceptual framework enabling users to: (1) re-play and share the stories in schools or at home, and (2) access the Website anytime and anywhere. Furthermore, the system stores the students work and activities over the system, allowing parents or teachers to monitor students’ works, and provide online feedback. The Website contains following main feature modules: Storytelling incorporates a variety of media such as audio, text and graphics in presenting the stories. It introduces the children to various kinds of traditional Amazigh oral narratives. The focus of this module is to project the positive values and images of stories using digital storytelling technique. Besides development good moral sense in children using projected positive images and moral values, it also allows children to practice their comprehending and listening skills. Reading module is developed based on multimedia material approach which offers the potential for addressing the challenges of reading instruction. This module is able to stimulate children and develop reading practice indirectly due to the tutoring strategies of scaffolding, self-explanation and hyperlinks offered in this module. Word Enhancement assists the children in understanding the story and appreciating the good moral values more efficiently. The difficult words or vocabularies are attached to present the explanation, which makes the children understand the vocabulary better. In conclusion, we believe that the interactive multimedia storytelling reveals an interesting and exciting tool for learning Amazigh. We plan to address some learning issues, in particularly the uses of activities to test and evaluate the children on their overall understanding of story and words presented in the learning modules.

Keywords: Amazigh language, e-learning, storytelling, language teaching

Procedia PDF Downloads 389

Authors: N. Manjunatha, K. T. Rangswamy, N. Nagaraju, H. A. Prameela, P. Rudraswamy, M. Krishnareddy

Abstract:

The identification of virus in cowpea especially potyviruses is confusing. Even though there are several studies on viruses causing diseases in cowpea, difficult to distinguish based on symptoms and serological detection. The differentiation of potyviruses considering as a constraint, the present study is initiated for molecular characterization, host plant resistance and epidemiology of the BCMV infecting cowpea. The etiological agent causing cowpea mosaic was identified as Bean Common Mosaic Virus (BCMV) on the basis of RT-PCR and electron microscopy. An approximately 750bp PCR product corresponding to coat protein (CP) region of the virus and the presence of long flexuous filamentous particles measuring about 952 nm in size typical to genus potyvirus were observed under electron microscope. The characterized virus isolate genome had 10054 nucleotides, excluding the 3’ terminal poly (A) tail. Comparison of polyprotein of the virus with other potyviruses showed similar genome organization with 9 cleavage sites resulted in 10 functional proteins. The pairwise sequence comparison of individual genes, P1 showed most divergent, but CP gene was less divergent at nucleotide and amino acid level. A phylogenetic tree constructed based on multiple sequence alignments of the polyprotein nucleotide and amino acid sequences of cowpea BCMV and potyviruses showed virus is closely related to BCMV-HB. Whereas, Soybean variant of china (KJ807806) and NL1 isolate (AY112735) showed 93.8 % (5’UTR) and 94.9 % (3’UTR) homology respectively with other BCMV isolates. This virus transmitted to different leguminous plant species and produced systemic symptoms under greenhouse conditions. Out of 100 cowpea genotypes screened, three genotypes viz., IC 8966, V 5 and IC 202806 showed immune reaction in both field and greenhouse conditions. Single marker analysis (SMA) was revealed out of 4 SSR markers linked to BCMV resistance, M135 marker explains 28.2 % of phenotypic variation (R2) and Polymorphic information content (PIC) value of these markers was ranged from 0.23 to 0.37. The correlation and regression analysis showed rainfall, and minimum temperature had significant negative impact and strong relationship with aphid population, whereas weak correlation was observed with disease incidence. Path coefficient analysis revealed most of the weather parameters exerted their indirect contributions to the aphid population and disease incidence except minimum temperature. This study helps to identify specific gaps in knowledge for researchers who may wish to further analyse the science behind complex interactions between vector-virus and host in relation to the environment. The resistant genotypes identified are could be effectively used in resistance breeding programme.

Keywords: cowpea, epidemiology, genotypes, virus

Procedia PDF Downloads 219

Authors: Yazan S. Khaled, Shazana Shamsuddin, Jim Tiernan, Mike McPherson, Thomas Hughes, Paul Millner, David G. Jayne

Abstract:

Introduction: There is a need for real-time imaging of colorectal cancer (CRC) to allow tailored surgery to the disease stage. Fluorescence guided laparoscopic imaging of primary colorectal cancer and the draining lymphatics would potentially bring stratified surgery into clinical practice and realign future CRC management to the needs of patients. Fluorescent nanoparticles can offer many advantages in terms of intra-operative imaging and therapy (theranostic) in comparison with traditional soluble reagents. Nanoparticles can be functionalised with diverse reagents and then targeted to the correct tissue using an antibody or Affimer (artificial binding protein). We aimed to develop and test fluorescent silica nanoparticles and targeted against CRC using an anti-carcinoembryonic antigen (CEA) Affimer (Aff). Methods: Anti-CEA and control Myoglobin Affimer binders were subcloned into the expressing vector pET11 followed by transformation into BL21 Star™ (DE3) E.coli. The expression of Affimer binders was induced using 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). Cells were harvested, lysed and purified using nickle chelating affinity chromatography. The photosensitiser Foslip (soluble analogue of 5,10,15,20-Tetra(m-hydroxyphenyl) chlorin) was incorporated into the core of silica nanoparticles using water-in-oil microemulsion technique. Anti-CEA or control Affs were conjugated to silica nanoparticles surface using sulfosuccinimidyl-4-(N-maleimidomethyl) cyclohexane-1-carboxylate (sulfo SMCC) chemical linker. Binding of CEA-Aff or control nanoparticles to colorectal cancer cells (LoVo, LS174T and HC116) was quantified in vitro using confocal microscopy. Results: The molecular weights of the obtained band of Affimers were ~12.5KDa while the diameter of functionalised silica nanoparticles was ~80nm. CEA-Affimer targeted nanoparticles demonstrated 9.4, 5.8 and 2.5 fold greater fluorescence than control in, LoVo, LS174T and HCT116 cells respectively (p < 0.002) for the single slice analysis. A similar pattern of successful CEA-targeted fluorescence was observed in the maximum image projection analysis, with CEA-targeted nanoparticles demonstrating 4.1, 2.9 and 2.4 fold greater fluorescence than control particles in LoVo, LS174T, and HCT116 cells respectively (p < 0.0002). There was no significant difference in fluorescence for CEA-Affimer vs. CEA-Antibody targeted nanoparticles. Conclusion: We are the first to demonstrate that Foslip-doped silica nanoparticles conjugated to anti-CEA Affimers via SMCC allowed tumour cell-specific fluorescent targeting in vitro, and had shown sufficient promise to justify testing in an animal model of colorectal cancer. CEA-Affimer appears to be a suitable targeting molecule to replace CEA-Antibody. Targeted silica nanoparticles loaded with Foslip photosensitiser is now being optimised to drive photodynamic killing, via reactive oxygen generation.

Keywords: colorectal cancer, silica nanoparticles, Affimers, antibodies, imaging

Procedia PDF Downloads 226

Authors: Dmitriy Babenko, Sergey Yegorov, Ilya Korshukov, Aidana Sultanbekova, Valentina Barkhanskaya, Tatiana Bashirova, Yerzhan Zhunusov, Yevgeniya Li, Viktoriya Parakhina, Svetlana Kolesnichenko, Yeldar Baiken, Aruzhan Pralieva, Zhibek Zhumadilova, Matthew S. Miller, Gonzalo H. Hortelano, Anar Turmuhambetova, Antonella E. Chesca, Irina Kadyrova

Abstract:

Introduction: The rapidly evolving COVID-19 pandemic, along with the re-emergence of pathogens causing acute respiratory infections (ARI), has necessitated the development of novel diagnostic tools to differentiate various causes of ARI. MALDI-MS, due to its wide usage and affordability, has been proposed as a potential instrument for diagnosing SARS-CoV-2 versus non-SARS-CoV-2 ARI. The aim of this study was to investigate the potential of MALDI-MS in conjunction with a machine learning model to accurately distinguish between symptomatic infections caused by SARS-CoV-2 and non-SARS-CoV-2 during both the early and later phases of the pandemic. Furthermore, this study aimed to analyze mass spectrometry (MS) data obtained from nasal swabs of healthy individuals. Methods: We gathered mass spectra from 252 samples, comprising 108 SARS-CoV-2-positive samples obtained in 2020 (Covid 2020), 7 SARS-CoV- 2-positive samples obtained in 2023 (Covid 2023), 71 samples from symptomatic individuals without SARS-CoV-2 (Control non-Covid ARVI), and 66 samples from healthy individuals (Control healthy). All the samples were subjected to RT-PCR testing. For data analysis, we employed the caret R package to train and test seven machine-learning algorithms: C5.0, KNN, NB, RF, SVM-L, SVM-R, and XGBoost. We conducted a training process using a five-fold (outer) nested repeated (five times) ten-fold (inner) cross-validation with a randomized stratified splitting approach. Results: In this study, we utilized the Covid 2020 dataset as a case group and the non-Covid ARVI dataset as a control group to train and test various machine learning (ML) models. Among these models, XGBoost and SVM-R demonstrated the highest performance, with accuracy values of 0.97 [0.93, 0.97] and 0.95 [0.95; 0.97], specificity values of 0.86 [0.71; 0.93] and 0.86 [0.79; 0.87], and sensitivity values of 0.984 [0.984; 1.000] and 1.000 [0.968; 1.000], respectively. When examining the Covid 2023 dataset, the Naive Bayes model achieved the highest classification accuracy of 43%, while XGBoost and SVM-R achieved accuracies of 14%. For the healthy control dataset, the accuracy of the models ranged from 0.27 [0.24; 0.32] for k-nearest neighbors to 0.44 [0.41; 0.45] for the Support Vector Machine with a radial basis function kernel. Conclusion: Therefore, ML models trained on MALDI MS of nasopharyngeal swabs obtained from patients with Covid during the initial phase of the pandemic, as well as symptomatic non-Covid individuals, showed excellent classification performance, which aligns with the results of previous studies. However, when applied to swabs from healthy individuals and a limited sample of patients with Covid in the late phase of the pandemic, ML models exhibited lower classification accuracy.

Keywords: SARS-CoV-2, MALDI-TOF MS, ML models, nasopharyngeal swabs, classification

Procedia PDF Downloads 93

Authors: María-Luisa Pinto-Salamanca, Wilson-Javier Pérez-Holguín

Abstract:

Reconstruction of contact forces is a fundamental technique for analyzing the properties of a touched object and is essential for regulating the grip force in slip control loops. This is based on the processing of the distribution, intensity, and direction of the forces during the capture of the sensors. Currently, efficient hardware alternatives have been used more frequently in different fields of application, allowing the implementation of computationally complex algorithms, as is the case with tactile signal processing. The use of hardware for smart tactile sensing systems is a research area that promises to improve the processing time and portability requirements of applications such as artificial skin and robotics, among others. The literature review shows that hardware implementations are present today in almost all stages of smart tactile detection systems except in the force reconstruction process, a stage in which they have been less applied. This work presents a hardware implementation of a model-driven reported in the literature for the contact force reconstruction of flat and rigid tactile sensor arrays from normal stress data. From the analysis of a software implementation of such a model, this implementation proposes the parallelization of tasks that facilitate the execution of matrix operations and a two-dimensional optimization function to obtain a vector force by each taxel in the array. This work seeks to take advantage of the parallel hardware characteristics of Field Programmable Gate Arrays, FPGAs, and the possibility of applying appropriate techniques for algorithms parallelization using as a guide the rules of generalization, efficiency, and scalability in the tactile decoding process and considering the low latency, low power consumption, and real-time execution as the main parameters of design. The results show a maximum estimation error of 32% in the tangential forces and 22% in the normal forces with respect to the simulation by the Finite Element Modeling (FEM) technique of Hertzian and non-Hertzian contact events, over sensor arrays of 10×10 taxels of different sizes. The hardware implementation was carried out on an MPSoC XCZU9EG-2FFVB1156 platform of Xilinx® that allows the reconstruction of force vectors following a scalable approach, from the information captured by means of tactile sensor arrays composed of up to 48 × 48 taxels that use various transduction technologies. The proposed implementation demonstrates a reduction in estimation time of x / 180 compared to software implementations. Despite the relatively high values of the estimation errors, the information provided by this implementation on the tangential and normal tractions and the triaxial reconstruction of forces allows to adequately reconstruct the tactile properties of the touched object, which are similar to those obtained in the software implementation and in the two FEM simulations taken as reference. Although errors could be reduced, the proposed implementation is useful for decoding contact forces for portable tactile sensing systems, thus helping to expand electronic skin applications in robotic and biomedical contexts.

Keywords: contact forces reconstruction, forces estimation, tactile sensor array, hardware implementation

Procedia PDF Downloads 185

Authors: Vivek Mahadev, Manoj Kumar, Neelu Mathur, Brahm Dutt Pandey

Abstract:

Perimeter guarding and protection of critical installations require prompt intrusion detection and assessment to take effective countermeasures. Currently, visual and electronic surveillance are the primary methods used for perimeter guarding. These methods can be costly and complicated, requiring careful planning according to the location and terrain. Moreover, these methods often struggle to detect stealthy and camouflaged insurgents. The object of the present work is to devise a surveillance technique using seismic sensors that overcomes the limitations of existing systems. The aim is to improve intrusion detection, assessment, and characterization by utilizing seismic sensors. Most of the similar systems have only two types of intrusion detection capability viz., human or vehicle. In our work we could even categorize further to identify types of intrusion activity such as walking, running, group walking, fence jumping, tunnel digging and vehicular movements. A virtual fence of 60 meters at GCNEP, Bahadurgarh, Haryana, India, was created by installing four underground geophones at a distance of 15 meters each. The signals received from these geophones are then processed to find unique seismic signatures called features. Various feature optimization and selection methodologies, such as LightGBM, Boruta, Random Forest, Logistics, Recursive Feature Elimination, Chi-2 and Pearson Ratio were used to identify the best features for training the machine learning models. The trained models were developed using algorithms such as supervised support vector machine (SVM) classifier, kNN, Decision Tree, Logistic Regression, Naïve Bayes, and Artificial Neural Networks. These models were then used to predict the category of events, employing weighted ensemble voting to analyze and combine their results. The models were trained with 1940 training events and results were evaluated with 831 test events. It was observed that using the weighted ensemble voting increased the efficiency of predictions. In this study we successfully developed and deployed the virtual fence using geophones. Since these sensors are passive, do not radiate any energy and are installed underground, it is impossible for intruders to locate and nullify them. Their flexibility, quick and easy installation, low costs, hidden deployment and unattended surveillance make such systems especially suitable for critical installations and remote facilities with difficult terrain. This work demonstrates the potential of utilizing seismic sensors for creating better perimeter guarding and protection systems using multiple machine learning models in weighted ensemble voting. In this study the virtual fence achieved an intruder detection efficiency of over 97%.

Keywords: geophone, seismic perimeter surveillance, machine learning, weighted ensemble method

Procedia PDF Downloads 66

Authors: Zahra Mohajer, Afsaneh Soltani

Abstract:

Doping is a major issue in competitive sports and is favored by vast groups of athletes. The feeling of being higher-ranking than others and gaining fame has caused many athletes to misuse drugs. The definition of doping is to use prohibited substances and/or methods that help physical or mental performances or both. Doping counts as the illegal use of chemical substances or drugs, excessive amounts of physiological substances to increase the performance at or out of competition or even the use of inappropriate medications to treat an injury to gain the ability to participate in a competition. The International Olympic Committee (IOC) and World Anti-Doping Agency (WADA) have forbidden these substances to ensure fair and equal competition and also the health of the competitors. As of 2004 WADA has published an international list of illegal substances used for doping, which is updated annually. In the process of the Genome Project scientists have gained the ability to treat numerous diseases by gene therapy, which may result in bodily performance increase and therefore a potential opportunity to misuse by some athletes. Gene doping is defined as the non-therapeutic direct and indirect genetic modifications using genetic materials that can improve the performances in sports events. Biosynthetic drugs are a form of indirect genetic engineering. The method can be performed in three ways such as injecting the DNA directly into the muscle, inserting the genetically engineered cells, or transferring the DNA using a virus as a vector. Erythropoietin is a hormone majorly released by the kidney and in small amounts by the liver. Its function is to stimulate the erythropoiesis and therefore the more production of red blood cells (RBC) which causes an increase in Hemoglobin (Hb). During this process, the oxygen delivery to muscles will increase, which will improve athletic performance and postpone exhaustion. There are ways to increase the oxygen transferred to muscles such as blood transfusion, stimulating the production of red blood cells by using Erythropoietin (EPO), and also using allosteric effectors of Hemoglobin. EPO can either be injected as a protein or can be inserted into the cells as the gene which encodes EPO. Adeno-associated viruses have been employed to deliver the EPO gene to the cells. Employing the genes that naturally exist in the human body such as the EPO gene can reduce the risk of detecting gene doping. The first research about blood doping was conducted in 1947. The study has shown that an increase in hematocrit (HCT) up to 55% following homologous transfusion makes it more unchallenging for the body to perform the exercise at the altitude. Thereafter athletes’ attraction to blood infusion escalated. Also, a study has demonstrated that by reinfusing their own blood 4 weeks after being drawn, three men have shown a rise in Hb level which improved the oxygen uptake, and a delay in exhaustion. The list of performance-enhancing drugs is published by WADA annually and includes the following drugs: anabolic agents, hormones, Beta-2 agonists, Beta-blockers, Diuretics, Stimulants, narcotics, cannabinoids, and corticosteroids.

Keywords: doping, PEDs, sports, WADA

Procedia PDF Downloads 101

Authors: Noura Al Akkari, Aurélie Foucquier, Sylvain Lespinats

Abstract:

Load monitoring is a management process for energy consumption towards energy savings and energy efficiency. Non Intrusive Load Monitoring (NILM) is one method of load monitoring used for disaggregation purposes. NILM is a technique for identifying individual appliances based on the analysis of the whole residence data retrieved from the main power meter of the house. Our NILM framework starts with data acquisition, followed by data preprocessing, then event detection, feature extraction, then general appliance modeling and identification at the final stage. The event detection stage is a core component of NILM process since event detection techniques lead to the extraction of appliance features. Appliance features are required for the accurate identification of the household devices. In this research work, we aim at developing a new event detection methodology with accurate load disaggregation to extract appliance features. Time-domain features extracted are used for tuning general appliance models for appliance identification and classification steps. We use unsupervised algorithms such as Dynamic Time Warping (DTW). The proposed method relies on detecting areas of operation of each residential appliance based on the power demand. Then, detecting the time at which each selected appliance changes its states. In order to fit with practical existing smart meters capabilities, we work on low sampling data with a frequency of (1/60) Hz. The data is simulated on Load Profile Generator software (LPG), which was not previously taken into consideration for NILM purposes in the literature. LPG is a numerical software that uses behaviour simulation of people inside the house to generate residential energy consumption data. The proposed event detection method targets low consumption loads that are difficult to detect. Also, it facilitates the extraction of specific features used for general appliance modeling. In addition to this, the identification process includes unsupervised techniques such as DTW. To our best knowledge, there exist few unsupervised techniques employed with low sampling data in comparison to the many supervised techniques used for such cases. We extract a power interval at which falls the operation of the selected appliance along with a time vector for the values delimiting the state transitions of the appliance. After this, appliance signatures are formed from extracted power, geometrical and statistical features. Afterwards, those formed signatures are used to tune general model types for appliances identification using unsupervised algorithms. This method is evaluated using both simulated data on LPG and real-time Reference Energy Disaggregation Dataset (REDD). For that, we compute performance metrics using confusion matrix based metrics, considering accuracy, precision, recall and error-rate. The performance analysis of our methodology is then compared with other detection techniques previously used in the literature review, such as detection techniques based on statistical variations and abrupt changes (Variance Sliding Window and Cumulative Sum).

Keywords: electrical disaggregation, DTW, general appliance modeling, event detection

Procedia PDF Downloads 64