Search results for: imbalanced multiclass classifi-cation

Authors: Srinidhi. A, Naveed Ahmed, Twinkle Hareendran, Vriksha Prakash

Abstract:

This work reviews some of the advanced image processing techniques for deep learning applications. Object detection by template matching, image denoising, edge detection, and super-resolution modelling are but a few of the tasks. The paper looks in into great detail, given that such tasks are crucial preprocessing steps that increase the quality and usability of image datasets in subsequent deep learning tasks. We review some of the methods for the assessment of image quality, more specifically sharpness, which is crucial to ensure a robust performance of models. Further, we will discuss the development of deep learning models specific to facial emotion detection, age classification, and gender classification, which essentially includes the preprocessing techniques interrelated with model performance. Conclusions from this study pinpoint the best practices in the preparation of image datasets, targeting the best trade-off between computational efficiency and retaining important image features critical for effective training of deep learning models.

Keywords: image processing, machine learning applications, template matching, emotion detection

Procedia PDF Downloads 17

Authors: Stefano Zappacosta, Cristiano Bove, Maria Carmela Marinelli, Paola di Lauro, Katarina Spasenovic, Lorenzo Ostano, Giuseppe Aiello, Marco Pietrosanto

Abstract:

This paper proposes a novel neural network approach for assessing flood hazard mapping. The core of the model is a machine learning component fed by frequency ratios, namely statistical correlations between flood event occurrences and a selected number of topographic properties. The proposed hybrid model can be used to classify four different increasing levels of hazard. The classification capability was compared with the flood hazard mapping River Basin Plans (PAI) designed by the Italian Institute for Environmental Research and Defence, ISPRA (Istituto Superiore per la Protezione e la Ricerca Ambientale). The study area of Piemonte, an Italian region, has been considered without loss of generality. The frequency ratios may be used as a standalone block to model the flood hazard mapping. Nevertheless, the mixture with a neural network improves the classification power of several percentage points, and may be proposed as a basic tool to model the flood hazard map in a wider scope.

Keywords: flood modeling, hazard map, neural networks, hydrogeological risk, flood risk assessment

Procedia PDF Downloads 178

Authors: Mukesh Singh Boori, Vít Voženílek

Abstract:

Olomouc is a unique and complex landmark with widespread forestation and land use. This research work was conducted to assess important and complex land use change trajectories in Olomouc region. Multi-temporal satellite data from 1991, 2001 and 2013 were used to extract land use/cover types by object oriented classification method. To achieve the objectives, three different aspects were used: (1) Calculate the quantity of each transition; (2) Allocate location based landscape pattern (3) Compare land use/cover evaluation procedure. Land cover change trajectories shows that 16.69% agriculture, 54.33% forest and 21.98% other areas (settlement, pasture and water-body) were stable in all three decade. Approximately 30% of the study area maintained as a same land cove type from 1991 to 2013. Here broad scale of political and socio-economic factors was also affect the rate and direction of landscape changes. Distance from the settlements was the most important predictor of land cover change trajectories. This showed that most of landscape trajectories were caused by socio-economic activities and mainly led to virtuous change on the ecological environment.

Keywords: remote sensing, land use/cover, change trajectories, image classification

Procedia PDF Downloads 404

Authors: Madhuri S. Ghate, Rahul P. Chavhan, Shriya S. Nahar

Abstract:

Learning objective: •To describe Magnetic Resonance Imaging (MRI) imaging appearances of typical and atypical uterine fibroids with emphasis on differentiating it from other similar conditions. •To classify uterine fibroids according to International Federation of Gynecology and Obstetrics (FIGO) Sub-classifications system and emphasis on its clinical significance. •To show cases with atypical imaging appearances atypical fibroids Material and methods: MRI of Pelvis had been performed in symptomatic women of child bearing age group on 1.5T and 3T MRI using T1, T2, STIR, FAT SAT, DWI sequences. Contrast was administered when degeneration was suspected. Imaging appearances of Atypical fibroids and various degenerations in fibroids were studied. Fibroids were classified using FIGO Sub-classification system. Its impact on surgical decision making and clinical outcome were also studied qualitatively. Results: Intramural fibroids were most common (14 patients), subserosal 7 patients, submucosal 5 patients . 6 patients were having multiple fibroids. 7 were having atypical fibroids. (1 hyaline degeneration, 1 cystic degeneration, 1 fatty, 1 necrosis and hemorrhage, 1 red degeneration, 1 calcification, 1 unusual large bilobed growth). Fibroids were classified using FIGO system. In uterus conservative surgeries, the lesser was the degree of myometrial invasion of fibroid, better was the fertility outcome. Conclusion: Relationship of fibroid with mucosal and serosal layers is important in the management of symptomatic fibroid cases. Risk to fertility involved in uterus conservative surgeries in women of child bearing age group depends on the extent of myometrial invasion of fibroids. FIGO system provides better insight into the degree of myometrial invasion. Knowledge about the atypical appearances of fibroids is important to avoid diagnostic confusion and untoward treatment.

Keywords: degeneration, FIGO sub-classification, MRI pelvis, uterine fibroids

Procedia PDF Downloads 92

Authors: Chakrit Thapphan, Suteeraporn Chaowattanapanit, Sorutsiri Chareonsudjai, Wisitsak Phoksawat, Supranee Phantanawiboon, Kiatichai Faksri, Steve W. Edwards, Kanin Salao

Abstract:

Background: Psoriasis is a chronic inflammatory disease of the skin that is mediated by crosstalk between keratinocytes and immune cells, especially CD4+ T helper (Th) cells. To date, psoriasis is established as a T helper 17 (Th17) cell-mediated inflammatory process driven by the over-expression of Th17. However, the role of other CD4+T helper cells is rather controversial. Objective: Our study, thereby, aimed to characterize and analyze T cell subsets in the circulating blood of psoriasis patients and compare them to healthy controls. Methods: Peripheral blood mononuclear cells were isolated from the participants and stained with fluorescent dye-conjugated monoclonal antibodies specific for intracellular cytokines, including interferon-gamma (IFN- γ), interleukin (IL-4), IL-17 and forkhead box P3 (FOXP3), that can be used to define T helper 1 (Th1) cells, T helper 2 (Th2), T helper 17 (Th17) and regulatory T cells (Treg) respectively. Results: We found that the numbers of Th2 (59.6% ± 17.0) and Th17 (4.0% ± 2.0) cells in the circulating blood of psoriasis patients were significantly higher than those of the healthy controls (p= 0.0007 and 0.0013 respectively). In contrast, the numbers of Th1 and Treg cells were not significantly different between psoriasis patients and healthy controls (p= 0.0593 and 0.8518, respectively). Additionally, when adjusting these numbers of Th cells to Treg, we observed a similar trend that the ratio of Th2/Treg and Th17/Treg also elevated (p = 0.0007 and 0.0047, respectively). Conclusion: Taken together, our results suggest an imbalanced T exhibit toward the Th2 and Th17 skewed-immune responses in psoriasis patients.

Keywords: psoriasis, Th cell subsets, Th2 cells, Th17 cells, Treg cells

Procedia PDF Downloads 77

Authors: Maha Shamseddine, Amjad Nusayr, Wassim Itani

Abstract:

In this paper, we present a Platform-as-a-Service (PaaS) model for controlling the privacy enforcement mechanisms applied on user data when stored and processed in Cloud data centers. The proposed architecture consists of establishing user configurable ‘sticky’ policies on the Graphical User Interface (GUI) data-bound components during the application development phase to specify the details of privacy enforcement on the contents of these components. Various privacy classification classes on the data components are formally defined to give the user full control on the degree and scope of privacy enforcement including the type of execution containers to process the data in the Cloud. This not only enhances the privacy-awareness of the developed Cloud services, but also results in major savings in performance and energy efficiency due to the fact that the privacy mechanisms are solely applied on sensitive data units and not on all the user content. The proposed design is implemented in a real PaaS cloud computing environment on the Microsoft Azure platform.

Keywords: privacy enforcement, platform-as-a-service privacy awareness, cloud computing privacy

Procedia PDF Downloads 227

Authors: Linghui Meng, James Atlas, Deborah Munro

Abstract:

There is an increasing demand for prosthetics capable of mimicking natural limb movements and hand gestures, but precise movement control of prosthetics using only electrode signals continues to be challenging. This study considers the implementation of machine learning as a means of improving accuracy and presents an initial investigation into hand gesture recognition using models based on electromyographic (EMG) signals. EMG signals, which capture muscle activity, are used as inputs to machine learning algorithms to improve prosthetic control accuracy, functionality and adaptivity. Using logistic regression, a machine learning classifier, this study evaluates the accuracy of classifying two hand gestures from the publicly available Ninapro dataset using two-time series feature extraction algorithms: Time Series Feature Extraction (TSFE) and Convolutional Neural Networks (CNNs). Trials were conducted using varying numbers of EMG channels from one to eight to determine the impact of channel quantity on classification accuracy. The results suggest that although both algorithms can successfully distinguish between hand gesture EMG signals, CNNs outperform TSFE in extracting useful information for both accuracy and computational efficiency. In addition, although more channels of EMG signals provide more useful information, they also require more complex and computationally intensive feature extractors and consequently do not perform as well as lower numbers of channels. The findings also underscore the potential of machine learning techniques in developing more effective and adaptive prosthetic control systems.

Keywords: EMG, machine learning, prosthetic control, electromyographic prosthetics, hand gesture classification, CNN, computational neural networks, TSFE, time series feature extraction, channel count, logistic regression, ninapro, classifiers

Procedia PDF Downloads 31

Authors: Zyad Elkhadir, Khalid Chougdali, Mohammed Benattou

Abstract:

Most of the existing intrusion detection systems works on quantitative network traffic data with many irrelevant and redundant features, which makes detection process more time’s consuming and inaccurate. A several feature extraction methods, such as linear discriminant analysis (LDA), have been proposed. However, LDA suffers from the small sample size (SSS) problem which occurs when the number of the training samples is small compared with the samples dimension. Hence, classical LDA cannot be applied directly for high dimensional data such as network traffic data. In this paper, we propose two solutions to solve SSS problem for LDA and apply them to a network IDS. The first method, reduce the original dimension data using principal component analysis (PCA) and then apply LDA. In the second solution, we propose to use the pseudo inverse to avoid singularity of within-class scatter matrix due to SSS problem. After that, the KNN algorithm is used for classification process. We have chosen two known datasets KDDcup99 and NSLKDD for testing the proposed approaches. Results showed that the classification accuracy of (PCA+LDA) method outperforms clearly the pseudo inverse LDA method when we have large training data.

Keywords: LDA, Pseudoinverse, PCA, IDS, NSL-KDD, KDDcup99

Procedia PDF Downloads 227

Authors: Atanu K Samanta, Asim Ali Khan

Abstract:

Due to the acquisition of huge amounts of brain tumor magnetic resonance images (MRI) in clinics, it is very difficult for radiologists to manually interpret and segment these images within a reasonable span of time. Computer-aided diagnosis (CAD) systems can enhance the diagnostic capabilities of radiologists and reduce the time required for accurate diagnosis. An intelligent computer-aided technique for automatic detection of a brain tumor through MRI is presented in this paper. The technique uses the following computational methods; the Level Set for segmentation of a brain tumor from other brain parts, extraction of features from this segmented tumor portion using gray level co-occurrence Matrix (GLCM), and the Artificial Neural Network (ANN) to classify brain tumor images according to their respective types. The entire work is carried out on 50 images having five types of brain tumor. The overall classification accuracy using this method is found to be 98% which is significantly good.

Keywords: brain tumor, computer-aided diagnostic (CAD) system, gray-level co-occurrence matrix (GLCM), tumor segmentation, level set method

Procedia PDF Downloads 513

Authors: M. Khaled Abduesslam, Mohammed Ali, Basher H. Alsdai, Muhammad Nizam Inayati

Abstract:

This paper presents the voltage problem location classification using performance of Least Squares Support Vector Machine (LS-SVM) and Learning Vector Quantization (LVQ) in electrical power system for proper voltage problem location implemented by IEEE 39 bus New-England. The data was collected from the time domain simulation by using Power System Analysis Toolbox (PSAT). Outputs from simulation data such as voltage, phase angle, real power and reactive power were taken as input to estimate voltage stability at particular buses based on Power Transfer Stability Index (PTSI).The simulation data was carried out on the IEEE 39 bus test system by considering load bus increased on the system. To verify of the proposed LS-SVM its performance was compared to Learning Vector Quantization (LVQ). The results showed that LS-SVM is faster and better as compared to LVQ. The results also demonstrated that the LS-SVM was estimated by 0% misclassification whereas LVQ had 7.69% misclassification.

Keywords: IEEE 39 bus, least squares support vector machine, learning vector quantization, voltage collapse

Procedia PDF Downloads 442

Authors: Gianne Bianca P. Manalo, Ernesto A. Martin, Vanessa V. Velasco

Abstract:

The laying performance, egg quality, egg classification, and income over feed cost of Improved Philippine Mallard duck (Itik Pinas) were examined as influenced by garlic powder in drinking water. A total of 48 ducks (42 females and 6 males) were used in the study. The ducks were allocated into two treatments - with garlic powder (GP) and without garlic powder (control) in drinking water. Each treatment had three replicates with eight ducks (7 females and 1 male) per replication. The results showed that there was a significant (P = 0.03) difference in average egg weight where higher values were attained by ducks with GP (77.67 g ± 0.64) than the control (75.64 g ± 0.43). The supplementation of garlic powder in drinking water, however, did not affect the egg production, feed intake, FCR, egg mass, livability, egg quality and egg classification. The Itik Pinas with GP in drinking water had numerically higher income over feed cost than those without. GP in drinking water can be considered in raising Itik Pinas. Further studies on increasing level of GP and long feeding duration also merit consideration to substantiate the findings.

Keywords: phytogenic, garlic powder, Itik-Pinas, egg weight, egg production

Procedia PDF Downloads 84

Authors: Damian Czubak

Abstract:

The Białowieża Forest is a very valuable natural area, included in the World Natural Heritage at UNESCO, where, due to infestation by the bark beetle (Ips typographus), norway spruce (Picea abies) have deteriorated. This catastrophic scenario led to an increase in fire danger. This was due to the occurrence of large amounts of dead wood and grass cover, as light penetrated to the bottom of the stands. These factors in a dry state are materials that favour the possibility of fire and the rapid spread of fire. One of the objectives of the study was to monitor the quantitative and qualitative changes of combustible material on the permanent decay plots of spruce stands from 2012-2022. In addition, the size of the area with highly flammable vegetation was monitored and a classification of the stands of the Białowieża Forest by flammability classes was made. The key factor that determines the potential fire hazard of a forest is combustible material. Primarily its type, quantity, moisture content, size and spatial structure. Based on the inventory data on the areas of forest districts in the Białowieża Forest, the average fire load and its changes over the years were calculated. The analysis was carried out taking into account the changes in the health status of the stands and sanitary operations. The quantitative and qualitative assessment of fallen timber and fire load of ground cover used the results of the 2019 and 2021 inventories. Approximately 9,000 circular plots were used for the study. An assessment was made of the amount of potential fuel, understood as ground cover vegetation and dead wood debris. In addition, monitoring of areas with vegetation that poses a high fire risk was conducted using data from 2019 and 2021. All sub-areas were inventoried where vegetation posing a specific fire hazard represented at least 10% of the area with species characteristic of that cover. In addition to the size of the area with fire-prone vegetation, a very important element is the size of the fire load on the indicated plots. On representative plots, the biomass of the land cover was measured on an area of 10 m2 and then the amount of biomass of each component was determined. The resulting element of variability of ground covers in stands was their flammability classification. The classification developed made it possible to track changes in the flammability classes of stands over the period covered by the measurements.

Keywords: classification, combustible material, flammable vegetation, Norway spruce

Procedia PDF Downloads 93

Authors: Royan Dawud Aldian, Endah Purwanti, Soegianto Soelistiono

Abstract:

In this research we have been developed an automatic investigation to classify normal children voice or autistic by using modern computation technology that is computation based on artificial neural network. The superiority of this computation technology is its capability on processing and saving data. In this research, digital voice features are gotten from the coefficient of linear-predictive coding with auto-correlation method and have been transformed in frequency domain using fast fourier transform, which used as input of artificial neural network in back-propagation method so that will make the difference between normal children and autistic automatically. The result of back-propagation method shows that successful classification capability for normal children voice experiment data is 100% whereas, for autistic children voice experiment data is 100%. The success rate using back-propagation classification system for the entire test data is 100%.

Keywords: autism, artificial neural network, backpropagation, linier predictive coding, fast fourier transform

Procedia PDF Downloads 461

Authors: Ameur Abdelkader, Abed Bouarfa Hafida

Abstract:

Since its inception, predictive analysis has revolutionized the IT industry through its robustness and decision-making facilities. It involves the application of a set of data processing techniques and algorithms in order to create predictive models. Its principle is based on finding relationships between explanatory variables and the predicted variables. Past occurrences are exploited to predict and to derive the unknown outcome. With the advent of big data, many studies have suggested the use of predictive analytics in order to process and analyze big data. Nevertheless, they have been curbed by the limits of classical methods of predictive analysis in case of a large amount of data. In fact, because of their volumes, their nature (semi or unstructured) and their variety, it is impossible to analyze efficiently big data via classical methods of predictive analysis. The authors attribute this weakness to the fact that predictive analysis algorithms do not allow the parallelization and distribution of calculation. In this paper, we propose to extend the predictive analysis algorithm, Classification And Regression Trees (CART), in order to adapt it for big data analysis. The major changes of this algorithm are presented and then a version of the extended algorithm is defined in order to make it applicable for a huge quantity of data.

Keywords: predictive analysis, big data, predictive analysis algorithms, CART algorithm

Procedia PDF Downloads 142

Authors: Alberto Piatto, Zaza Nadja Lee Hansen, Peter Jacobsen

Abstract:

This paper examines the quantitative and qualitative factors influencing the buyer-supplier relationship. Understanding and acting on the right factors influencing supplier relationship management is crucial when a company outsource an important part of its business as it can be for engineering to order (ETO) company executing only the designing part in-house. Acting on these factors increase the quality of the relationship obtaining for both parties what they want and expect from an improved relationship. Best practices in supplier relationship management are considered and a case study of a large global company, called Cement A/S, operating in the cement business is carried out. One study is conducted including a large international company and hundreds of its suppliers. Data from the company is collected using semi-structured interviews and data from the suppliers is collected using a survey. Based on these inputs and an extensive literature review a classification of factors influencing the relationship buyer-supplier is presented and discussed. The results show that different managers among the company are assessing supplier from various perspectives, a standard approach to measure the performance of suppliers does not exist. The factors used nowadays in the company to measure performances of the suppliers are mostly related to time and cost. Quality is a key factor, but it has not been addressed properly since no data are available in the system. From a practical perspective, managers can learn from this paper which factors to consider when applying best practices of Supplier Relationship Management. Furthermore, from a theoretical perspective, this paper contributes with new knowledge in the area as limited research in collaboration with the company has been conducted. For this reason, a company, its suppliers and few studies for this type of industry have been conducted. For further research, it is suggested to define the correlation of factors to the profitability of the company and calculate its impact. When conducting this analysis it is important to focus on the efficient and effective use of factors that can be measurable and accepted from the supplier.

Keywords: buyer-supplier relationship, cement industry, classification of factors, ETO

Procedia PDF Downloads 283

Authors: Pablo Martinez, Emad Mohamed, Osama Mohsen, Yasser Mohamed

Abstract:

Highway and bridge agencies, such as the Ministry of Transportation in Ontario, use the Bridge Condition Index (BCI) which is defined as the weighted condition of all bridge elements to determine the rehabilitation priorities for its bridges. Therefore, accurate forecasting of BCI is essential for bridge rehabilitation budgeting planning. The large amount of data available in regard to bridge conditions for several years dictate utilizing traditional mathematical models as infeasible analysis methods. This research study focuses on investigating different classification models that are developed to predict the bridge condition index in the province of Ontario, Canada based on the publicly available data for 2800 bridges over a period of more than 10 years. The data preparation is a key factor to develop acceptable classification models even with the simplest one, the k-NN model. All the models were tested, compared and statistically validated via cross validation and t-test. A simple k-NN model showed reasonable results (within 0.5% relative error) when predicting the bridge condition in an incoming year.

Keywords: asset management, bridge condition index, data mining, forecasting, infrastructure, knowledge discovery in databases, maintenance, predictive models

Procedia PDF Downloads 191

Authors: Henni Sid Ahmed, Belbachir Mohamed Faouzi, Jean Caelen

Abstract:

Behavior recognition has been studied for realizing drivers assisting system and automated navigation and is an important studied field in the intelligent Building. In this paper, a recognition method of behavior recognition separated from a real image was studied. Images were divided into several categories according to the actual weather, distance and angle of view etc. SIFT was firstly used to detect key points and describe them because the SIFT (Scale Invariant Feature Transform) features were invariant to image scale and rotation and were robust to changes in the viewpoint and illumination. My goal is to develop a robust and reliable system which is composed of two fixed cameras in every room of intelligent building which are connected to a computer for acquisition of video sequences, with a program using these video sequences as inputs, we use SIFT represented different images of video sequences, and SVM (support vector machine) Lights as a programming tool for classification of images in order to classify people’s behaviors in the intelligent building in order to give maximum comfort with optimized energy consumption.

Keywords: video analysis, people behavior, intelligent building, classification

Procedia PDF Downloads 378

Authors: Ramón Aparicio-García, Gustavo Juárez Gracia, Jesús Álvarez Cedillo

Abstract:

A research line of the computer science that involve the study of the Human-Computer Interaction (HCI), which search to recognize and interpret the user intent by the storage and the subsequent analysis of the electrical signals of the brain, for using them in the control of electronic devices. On the other hand, the affective computing research applies the human emotions in the HCI process helping to reduce the user frustration. This paper shows the results obtained during the hardware and software development of a Brain Computer Interface (BCI) capable of recognizing the human emotions through the association of the brain electrical activity patterns. The hardware involves the sensing stage and analogical-digital conversion. The interface software involves algorithms for pre-processing of the signal in time and frequency analysis and the classification of patterns associated with the electrical brain activity. The methods used for the analysis and classification of the signal have been tested separately, by using a database that is accessible to the public, besides to a comparison among classifiers in order to know the best performing.

Keywords: affective computing, interface, brain, intelligent interaction

Procedia PDF Downloads 389

Authors: Ülle Parm, Kata Pedamäe, Jaak Jürimäe, Evelin Lätt, Aivar Orav, Anna-Liisa Tamm

Abstract:

Vegetarianism has gained popularity across the world. It`s being chosen for multiple reasons, but among Estonians, these have remained unknown. Previously, attention to bone health and probable nutrient deficiency of vegetarians has been paid and in vegetarians lower body mass index (BMI) and blood cholesterol level has been found but the results are inconclusive. The goal was to explain reasons for choosing vegetarian diet in Estonia and impact of vegetarianism to body composition – BMI, fat percentage (fat%), fat mass (FM), and fat free mass (FFM). The study group comprised of 68 vegetarians and 103 omnivorous. The determining body composition with DXA (Hologic) was concluded in 2013. Body mass (medical electronic scale, A&D Instruments, Abingdon, UK) and height (Martin metal anthropometer to the nearest 0.1 cm) were measured and BMI calculated (kg/m2). General data (physical activity level included) was collected with questionnaires. The main reasons why vegetarianism was chosen were the healthiness of the vegetarian diet (59%) and the wish to fight for animal rights (72%) Food additives were consumed by less than half of vegetarians, more often by men. Vegetarians had lower BMI than omnivores, especially amongst men. Based on BMI classification, vegetarians were less obese than omnivores. However, there were no differences in the FM, FFM and fat percentage figures of the two groups. Higher BMI might be the cause of higher physical activity level among omnivores compared with vegetarians. For classifying people as underweight, normal weight, overweight and obese both BMI and fat% criteria were used. By BMI classification in comparison with fat%, more people in the normal weight group were considered; by using fat% in comparison with BMI classification, however, more people categorized as overweight. It can be concluded that the main reasons for vegetarianism chosen in Estonia are healthiness of the vegetarian diet and the wish to fight for animal rights and vegetarian diet has no effect on body fat percentage, FM and FFM.

Keywords: body composition, body fat percentage, body mass index, vegetarianism

Procedia PDF Downloads 416

Authors: A. M. John-Otumu, M. M. Rahman, O. C. Nwokonkwo, M. C. Onuoha

Abstract:

Online social media networks have long served as a primary arena for group conversations, gossip, text-based information sharing and distribution. The use of natural language processing techniques for text classification and unbiased decision-making has not been far-fetched. Proper classification of this textual information in a given context has also been very difficult. As a result, we decided to conduct a systematic review of previous literature on sentiment classification and AI-based techniques that have been used in order to gain a better understanding of the process of designing and developing a robust and more accurate sentiment classifier that can correctly classify social media textual information of a given context between hate speech and inverted compliments with a high level of accuracy by assessing different artificial intelligence techniques. We evaluated over 250 articles from digital sources like ScienceDirect, ACM, Google Scholar, and IEEE Xplore and whittled down the number of research to 31. Findings revealed that Deep learning approaches such as CNN, RNN, BERT, and LSTM outperformed various machine learning techniques in terms of performance accuracy. A large dataset is also necessary for developing a robust sentiment classifier and can be obtained from places like Twitter, movie reviews, Kaggle, SST, and SemEval Task4. Hybrid Deep Learning techniques like CNN+LSTM, CNN+GRU, CNN+BERT outperformed single Deep Learning techniques and machine learning techniques. Python programming language outperformed Java programming language in terms of sentiment analyzer development due to its simplicity and AI-based library functionalities. Based on some of the important findings from this study, we made a recommendation for future research.

Keywords: artificial intelligence, natural language processing, sentiment analysis, social network, text

Procedia PDF Downloads 115

Authors: Mahesh Balan U, Rohith Srinivaas Mohanakrishnan, Venkat Subramanian

Abstract:

Positive and unlabeled learning (PU) has gained more attention in both academic and industry research literature recently because of its relevance to existing business problems today. Yet, there still seems to be some existing challenges in terms of validating the performance of PU learning, as the actual truth of unlabeled data points is still unknown in contrast to a binary classification where we know the truth. In this study, we propose a novel PU learning technique based on the Pseudo-F measure, where we address this research gap. In this approach, we train the PU model to discriminate the probability distribution of the positive and unlabeled in the validation and spy data. The predicted probabilities of the PU model have a two-fold validation – (a) the predicted probabilities of reliable positives and predicted positives should be from the same distribution; (b) the predicted probabilities of predicted positives and predicted unlabeled should be from a different distribution. We experimented with this approach on a credit marketing case study in one of the world’s biggest fintech platforms and found evidence for benchmarking performance and backtested using historical data. This study contributes to the existing literature on semi-supervised learning.

Keywords: PU learning, semi-supervised learning, pseudo f measure, classification

Procedia PDF Downloads 235

Authors: P. Senthil Kumari

Abstract:

Main concerns in data mining investigation are social controls of data mining for handling ambiguity, noise, or incompleteness on text data. We describe an innovative approach for unplanned text data detection of community networks achieved by classification mechanism. In a tangible domain claim with humble secrecy backgrounds provided by community network for evading annoying content is presented on consumer message partition. To avoid this, mining methodology provides the capability to unswervingly switch the messages and similarly recover the superiority of ordering. Here we designated learning-centered mining approaches with pre-processing technique to complete this effort. Our involvement of work compact with rule-based personalization for automatic text categorization which was appropriate in many dissimilar frameworks and offers tolerance value for permits the background of comments conferring to a variety of conditions associated with the policy or rule arrangements processed by learning algorithm. Remarkably, we find that the choice of classifier has predicted the class labels for control of the inadequate documents on community network with great value of effect.

Keywords: text mining, data classification, community network, learning algorithm

Procedia PDF Downloads 508

Authors: G. C. Tikkiwal, Mukesh Upadhyay

Abstract:

During the surveillance operations at war or peace time, the Radar operator gets a scatter of targets over the screen. This may be a tracked vehicle like tank vis-à-vis T72, BMP etc, or it may be a wheeled vehicle like ALS, TATRA, 2.5Tonne, Shaktiman or moving the army, moving convoys etc. The radar operator selects one of the promising targets into single target tracking (STT) mode. Once the target is locked, the operator gets a typical audible signal into his headphones. With reference to the gained experience and training over the time, the operator then identifies the random target. But this process is cumbersome and is solely dependent on the skills of the operator, thus may lead to misclassification of the object. In this paper, we present a technique using mathematical and statistical methods like fast fourier transformation (FFT) and principal component analysis (PCA) to identify the random objects. The process of classification is based on transforming the audible signature of target into music octave-notes. The whole methodology is then automated by developing suitable software. This automation increases the efficiency of identification of the random target by reducing the chances of misclassification. This whole study is based on live data.

Keywords: radar target, FFT, principal component analysis, eigenvector, octave-notes, DSP

Procedia PDF Downloads 394

Authors: Khawlah Al-Khalidi

Abstract:

Prevalence of lower third molar impactions and angulations among Yemeni population The purpose of this study was to look into the prevalence of lower third molars in a sample of patients from Ibb University Affiliated Hospital, as well as to study and categorise their position by using Pell and Gregory classification, and to look into a possible correlation between their position and the indication for extraction. Materials and methods: This is a retrospective, observational study in which a sample of 200 patients from Ibb University Affiliated Hospital were studied, including patient record validation and orthopantomography performed in screening appointments in people aged 16 to 21. Results and discussion: Males make up 63% of the sample, while people aged 19 to 20 make up 41.2%. Lower third molars were found in 365 of the 365 instances examined, accounting for 91% of the sample under study. According to Pell and Gregory's categorisation, the most common position is IIB, with 37%, followed by IIA with 21%; less common classes are IIIA, IC, and IIIC, with 1%, 3%, and 3%, respectively. It was feasible to determine that 56% of the lower third molars in the sample were recommended for extraction during the screening consultation. Finally, there are differences in third molar location and angulation. There was, however, a link between the available space for third molar eruption and the need for tooth extraction.

Keywords: lower third molar, extraction, Pell and Gregory classification, lower third molar impaction

Procedia PDF Downloads 55

Authors: Mahdieh Khalilinezhad, Silvana Dellepiane, Gianni Vernazza

Abstract:

The aim of the present work is to build a model based on tissue characterization that is able to discriminate pathological and non-pathological regions from three-phasic CT images. Based on feature selection in different phases, in this research, we design a neural network system that has optimal neuron number in a hidden layer. Our approach consists of three steps: feature selection, feature reduction, and classification. For each ROI, 6 distinct set of texture features are extracted such as first order histogram parameters, absolute gradient, run-length matrix, co-occurrence matrix, autoregressive model, and wavelet, for a total of 270 texture features. We show that with the injection of liquid and the analysis of more phases the high relevant features in each region changed. Our results show that for detecting HCC tumor phase3 is the best one in most of the features that we apply to the classification algorithm. The percentage of detection between these two classes according to our method, relates to first order histogram parameters with the accuracy of 85% in phase 1, 95% phase 2, and 95% in phase 3.

Keywords: multi-phasic liver images, texture analysis, neural network, hidden layer

Procedia PDF Downloads 262

Authors: Na Li, Yunwei Zhang, Yuhong Niu

Abstract:

The study discussed the patent protections of health system and enterprises in Shanghai. The comparisons of technical distribution and scopes of patent protections between Shanghai health system and enterprises were used by the methods of IPC classification, co-words analysis and visual social network. Results reflected a decreasing order within IPC A61 area, namely A61B, A61K, A61M, and A61F. A61B required to be further investigated. The highest authorized patents A61B17 of A61B of IPC A61 area was found. Within A61B17, fracture fixation, ligament reconstruction, cardiac surgery, and biopsy detection were regarded as common concerned fields by Shanghai health system and enterprises. However, compared with cardiac closure which Shanghai enterprises paid attention to, Shanghai health system was more inclined to blockages and hemostatic tools. The results also revealed that the scopes of patent protections of Shanghai enterprises were relatively centralized. Shanghai enterprises had a series of comprehensive strategies for protecting core patents. In contrast, Shanghai health system was considered to be lack of strategic patent protections for core patents.

Keywords: co-words analysis, IPC classification, patent protection, technical distribution

Procedia PDF Downloads 134

Authors: Ayse Pekrioglu Balkis, Sawash Mecid

Abstract:

Due to continued growth and rapid development of road construction in worldwide, road sub-layers consist of soil layers, therefore, identification and recognition of type of soil and soil behavior in different condition help to us to select soil according to specification and engineering characteristic, also if necessary sometimes stabilize the soil and treat undesirable properties of soils by adding materials such as bitumen, lime, cement, etc. If the soil beneath the road is not done according to the standards and construction will need more construction time. In this case, a large part of soil should be removed, transported and sometimes deposited. Then purchased sand and gravel is transported to the site and full depth filled and compacted. Stabilization by cement or other treats gives an opportunity to use the existing soil as a base material instead of removing it and purchasing and transporting better fill materials. Classification of soil according to AASHTOO system and USCS help engineers to anticipate soil behavior and select best treatment method. In this study soil classification and the relation between soil classification and stabilization method is discussed, cement stabilization with different percentages have been selected for soil treatment based on NCHRP. There are different parameters to define the strength of soil. In this study, CBR will be used to define the strength of soil. Cement by percentages, 0%, 3%, 7% and 10% added to soil for evaluation effect of added cement to CBR of treated soil. Implementation of stabilization process by different cement content help engineers to select an economic cement amount for the stabilization process according to project specification and characteristics. Stabilization process in optimum moisture content (OMC) and mixing rate effect on the strength of soil in the laboratory and field construction operation have been performed to see the improvement rate in strength and plasticity. Cement stabilization is quicker than a universal method such as removing and changing field soils. Cement addition increases CBR values of different soil types by the range of 22-69%.

Keywords: California Bearing Ratio, cement stabilization, clayey soil, mechanical properties

Procedia PDF Downloads 397

Authors: Naman Solanki, Souraj Mondal

Abstract:

With the world moving towards online communication, the video datastore has exploded in the past few years. Consequently, it has become crucial to analyse participant’s engagement levels in online communication videos. Engagement prediction of people in videos can be useful in many domains, like education, client meetings, dating, etc. Video-level or frame-level prediction of engagement for a user involves the development of robust models that can capture facial micro-emotions efficiently. For the development of an engagement prediction model, it is necessary to have a widely-accepted standard dataset for engagement analysis. DAiSEE is one of the datasets which consist of in-the-wild data and has a gold standard annotation for engagement prediction. Earlier research done using the DAiSEE dataset involved training and testing standard models like CNN-based models, but the results were not satisfactory according to industry standards. In this paper, a multi-level classification approach has been introduced to create a more robust model for engagement analysis using the DAiSEE dataset. This approach has recorded testing accuracies of 0.638, 0.7728, 0.8195, and 0.866 for predicting boredom level, engagement level, confusion level, and frustration level, respectively.

Keywords: computer vision, engagement prediction, deep learning, multi-level classification

Procedia PDF Downloads 114

Authors: Ermias A. Tegegn, Million Meshesha

Abstract:

Human Immunodeficiency Virus and Acquired Immunodeficiency Syndrome (HIV/AIDS) is a major cause of death for most African countries. Ethiopia is one of the seriously affected countries in sub Saharan Africa. Previously in Ethiopia, having HIV/AIDS was almost equivalent to a death sentence. With the introduction of Antiretroviral Therapy (ART), HIV/AIDS has become chronic, but manageable disease. The study focused on a data mining technique to predict future living status of HIV/AIDS patients at the time of drug regimen change when the patients become toxic to the currently taking ART drug combination. The data is taken from University of Gondar Hospital ART program database. Hybrid methodology is followed to explore the application of data mining on ART program dataset. Data cleaning, handling missing values and data transformation were used for preprocessing the data. WEKA 3.7.9 data mining tools, classification algorithms, and expertise are utilized as means to address the research problem. By using four different classification algorithms, (i.e., J48 Classifier, PART rule induction, Naïve Bayes and Neural network) and by adjusting their parameters thirty-two models were built on the pre-processed University of Gondar ART program dataset. The performances of the models were evaluated using the standard metrics of accuracy, precision, recall, and F-measure. The most effective model to predict the status of HIV patients with drug regimen substitution is pruned J48 decision tree with a classification accuracy of 98.01%. This study extracts interesting attributes such as Ever taking Cotrim, Ever taking TbRx, CD4 count, Age, Weight, and Gender so as to predict the status of drug regimen substitution. The outcome of this study can be used as an assistant tool for the clinician to help them make more appropriate drug regimen substitution. Future research directions are forwarded to come up with an applicable system in the area of the study.

Keywords: HIV drug regimen, data mining, hybrid methodology, predictive model

Procedia PDF Downloads 142

Authors: Hee-Jung Seo, Yunjung Lee, Areum Han, Heeyoung Park, Se-Hyuk Park

Abstract:

The purpose of this study was to analyze the relationship between happiness and depression among people with disabilities and to analyze the social phenomenon of leisure activities among them to promote physical and leisure activities for people with disabilities. The research methods included analyzing differences in happiness according to depression classification. A total of 281 people with disabilities were analyzed using SPSS WIN Ver. 29.0. In addition, the SumTrend platform was used to analyze terms related to 'leisure activities for the disabled.' The findings can be summarized into two main points: First, there were significant differences in happiness according to depression classification. Second, there were 20 mentions before COVID-19, 34 mentions after COVID-19, and currently 43 mentions, with high positive rates observed in each period. Based on these results, the following conclusions were drawn: First, measures for people with disabilities include strengthening online resources and services, social distancing response policies, improving accessibility, and providing support and financial assistance. Second, measures for non-disabled individuals emphasize the need for education and information provision, promoting dialogue and interaction, ensuring accessibility, and promoting inclusive cultural awareness and attitude change.

Keywords: leisure activities, individuals with disabilities, COVID-19 pandemic, depression

Procedia PDF Downloads 48