Search results for: Twitter data clustering

Authors: Eman AbuKhousa, Marwan Z. Bataineh

Abstract:

The 21^st century higher education and globalization challenge new faculty members to build effective professional networks and partnership with industry in order to accelerate their growth and success. This creates the need for community of practice (CoP)-oriented development approaches that focus on cognitive apprenticeship while considering individual predisposition and future career needs. This work adopts data mining, clustering analysis, and social networking technologies to present the CoP-Network as a virtual space that connects together similar career-aspiration individuals who are socially influenced to join and engage in a process for domain-related knowledge and practice acquisitions. The CoP-Network model can be integrated into higher education to extend traditional graduate and professional development programs.

Keywords: clustering analysis, community of practice, data mining, higher education, new faculty challenges, social network, social influence, professional development

Procedia PDF Downloads 163

Authors: Mai Alshareef

Abstract:

The Saudi government considers the modernisation of the education system as a critical component of the national development plan, Saudi Vision 2030; however, this sudden reform creates tension amongst Saudis. This study examines first the reflection of the social and digital education reform on stakeholders and the general Saudi public, and second, the influence of information and communication technologies (ICTs) on the ethnoreligious conflict in Saudi Arabia. This study employs Cyberconflict theory to examine conflicts in the real world and cyberspace. The findings are based on a qualitative case study methodology that uses netnography, an analysis of 3,750 Twitter posts and semi-structural interviews with 30 individuals, including key actors in the Saudi education sector and Twitter activists during 2019\2020. The methods utilised are guided by thematic analysis to map an understanding of factors that influence societal conflicts in Saudi Arabia, which in this case include religious, national, and gender identity. Elements of Cyberconflict theory are used to better understand how conflicting groups build their identities in connection to their ethnic/religious/cultural differences and competing national identities. The findings correspond to the ethnoreligious components of the Cyberconflict theory. Twitter became a battleground for liberals, conservatives, the Saudi public and elites, and it is used in a novel way to influence public opinion and to challenge the media monopoly. Opposing groups relied heavily on a discourse of exclusion and inclusion and showed ethnic and religious affiliations, national identity, and chauvinism. The findings add to existing knowledge in the cyberconflict field of study, and they also reveal outcomes that are critical to the Saudi Arabian national context.

Keywords: education, cyberconflict, Twitter, national identity

Procedia PDF Downloads 158

Authors: Sangeeta Yadav, Mantosh Biswas

Abstract:

In this paper, we proposed an improved color based K-mean algorithm for clustering of satellite Image (SAR). Our method comprises of two stages. The first step is an interactive selection process where users are required to input the number of colors (ncolor), number of clusters, and then they are prompted to select the points in each color cluster. In the second step these points are given as input to K-mean clustering algorithm that clusters the image based on color and Minimum Square Euclidean distance. The proposed method reduces the mixed pixel problem to a great extent.

Keywords: cluster, ncolor method, K-mean method, interactive selection process

Procedia PDF Downloads 279

Authors: Vani Singhal, Jitendra Parmar, Satyendra Singh Chouhan

Abstract:

Unseen class discovery has now become an important part of a machine-learning algorithm to judge new classes. Unseen classes are the classes on which the machine learning model is not trained on. With the advancement in technology and AI replacing humans, the amount of data has increased to the next level. So while implementing a model on real-world examples, we come across unseen new classes. Our aim is to find the number of unseen classes by using a hierarchical-based active learning algorithm. The algorithm is based on hierarchical clustering as well as active sampling. The number of clusters that we will get in the end will give the number of unseen classes. The total clusters will also contain some clusters that have unseen classes. Instead of first discovering unseen classes and then finding their number, we directly calculated the number by applying the algorithm. The dataset used is for intent classification. The target data is the intent of the corresponding query. We conclude that when the machine learning model will encounter real-world data, it will automatically find the number of unseen classes. In the future, our next work would be to label these unseen classes correctly.

Keywords: active sampling, hierarchical clustering, open world learning, unseen class discovery

Procedia PDF Downloads 150

Authors: William Wong Xiu Shun, Yoonjin Hyun, Mingyu Kim, Seongi Choi, Namgyu Kim

Abstract:

Recently, the demand of extracting the R&D keywords from the issues and using them in retrieving R&D information is increasing rapidly. But it is hard to identify the related issues or to distinguish them. Although the similarity between the issues cannot be identified, but with the R&D lexicon, the issues that always shared the same R&D keywords can be determined. In details, the R&D keywords that associated with particular issue is implied the key technology elements that needed to solve the problem of the particular issue. Furthermore, the related issues that sharing the same R&D keywords can be showed in a more systematic way through the issue clustering constructed from the perspective of R&D. Thus, sharing of the R&D result and reusable of the R&D technology can be facilitated. Indirectly, the redundancy of investment on the same R&D can be reduce as the R&D information can be shared between those corresponding issues and reusability of the related R&D can be improved. Therefore, a methodology of constructing an issue clustering from the perspective of common R&D keywords is proposed to satisfy the demands mentioned.

Keywords: clustering, social network analysis, text mining, topic analysis

Procedia PDF Downloads 557

Authors: Anuj Abraham, N. Pappa, Daniel Honc, Rahul Sharma

Abstract:

In this paper, an analysis of some model order reduction techniques is presented. A new hybrid algorithm for model order reduction of linear time invariant systems is compared with the conventional techniques namely Balanced Truncation, Hankel Norm reduction and Dominant Pole Algorithm (DPA). The proposed hybrid algorithm is known as Clustering Dominant Pole Algorithm (CDPA) is able to compute the full set of dominant poles and its cluster center efficiently. The dominant poles of a transfer function are specific eigenvalues of the state space matrix of the corresponding dynamical system. The effectiveness of this novel technique is shown through the simulation results.

Keywords: balanced truncation, clustering, dominant pole, Hankel norm, model reduction

Procedia PDF Downloads 581

Authors: Shanhua Hu

Abstract:

Social media as a critical propagandizing form of film, video games, and digital products has received substantial research attention, but there exists several critical barriers such as: (1) few studies exploring the internal and external connections of a product as part of the multimodal context that gives rise to readability and commercial return; (2) the lack of study of multimodal analysis in product’s official account of game publishers and its impact on users’ behaviors including purchase intention, social media engagement, and playing time; (3) no standardized ecologically-valid, game type-varying data can be used to study the complexity of official account’s postings within a time period. This proposed research helps to tackle these limitations in order to develop a model of readability study that is more ecologically valid, robust, and thorough. To accomplish this objective, this paper provides a more diverse dataset comprising different visual elements and messages collected from the official Twitter accounts of the Top 20 best-selling games of 2021. Video game companies target potential users through social media, a popular approach is to set up an official account to maintain exposure. Typically, major game publishers would create an official account on Twitter months before the game's release date to update on the game's development, announce collaborations, and reveal spoilers. Analyses of tweets from those official Twitter accounts would assist publishers and marketers in identifying how to efficiently and precisely deploy advertising to increase game sales. The purpose of this research is to determine how official game accounts use Twitter to attract new customers, specifically which types of messages are most effective at increasing sales. The dataset includes the number of days until the actual release date on Twitter posts, the readability of the post (Flesch Reading Ease Score, FRES), the number of emojis used, the number of hashtags, the number of followers of the mentioned users, the categorization of the posts (i.e., spoilers, collaborations, promotions), and the number of video views. The timeline of Twitter postings from official accounts will be compared to the history of pre-orders and sales figures to determine the potential impact of social media posts. This study aims to determine how the above-mentioned characteristics of official accounts' Twitter postings influence the sales of the game and to examine the possible causes of this influence. The outcome will provide researchers with a list of potential aspects that could influence people's judgments in limited-word posts. With the increased average online time, users would adapt more quickly than before in online information exchange and readings, such as the word to use sentence length, and the use of emojis or hashtags. The study on the promotion of official game accounts will not only enable publishers to create more effective promotion techniques in the future but also provide ideas for future research on the influence of social media posts with a limited number of words on consumers' purchasing decisions. Future research can focus on more specific linguistic aspects, such as precise word choice in advertising.

Keywords: engagement, official account, promotion, twitter, video game

Procedia PDF Downloads 56

Authors: Caio Marcelo Nunes, Anderson da Silva Soares, Gustavo Teodoro Laureano, Clarimar Jose Coelho

Abstract:

This paper presents a methodology to pixel recognition in aerial images using fuzzy $c$-means algorithm. This algorithm is a alternative to recognize areas considering uncertainties and inaccuracies. Traditional clustering technics are used in recognizing of multispectral images of earth's surface. This technics recognize well-defined borders that can be easily discretized. However, in the real world there are many areas with uncertainties and inaccuracies which can be mapped by clustering algorithms that use fuzzy sets. The methodology presents in this work is applied to multispectral images obtained from Landsat-5/TM satellite. The pixels are joined using the $c$-means algorithm. After, a classification process identify the types of surface according the patterns obtained from spectral response of image surface. The classes considered are, exposed soil, moist soil, vegetation, turbid water and clean water. The results obtained shows that the fuzzy clustering identify the real type of the earth's surface.

Keywords: aerial images, fuzzy clustering, image processing, pattern recognition

Procedia PDF Downloads 451

Authors: Maryam Maleki, Esther Mead, Mohammad Arani, Nitin Agarwal

Abstract:

This study is conducted to examine how opposing viewpoints related to COVID-19 were diffused on Twitter. To accomplish this, six datasets using two epidemiological models, SIR (Susceptible, Infected, Recovered) and SEIZ (Susceptible, Exposed, Infected, Skeptics), were analyzed. The six datasets were chosen because they represent opposing viewpoints on the COVID-19 pandemic. Three of the datasets contain anti-subject hashtags, while the other three contain pro-subject hashtags. The time frame for all datasets is three years, starting from January 2020 to December 2022. The findings revealed that while both models were effective in evaluating the propagation trends of these polarizing viewpoints, the SEIZ model was more accurate with a relatively lower error rate (6.7%) compared to the SIR model (17.3%). Additionally, the relative error for both models was lower for anti-subject hashtags compared to pro-subject hashtags. By leveraging epidemiological models, insights into the propagation trends of polarizing viewpoints on Twitter were gained. This study paves the way for the development of methods to prevent the spread of ideas that lack scientific evidence while promoting the dissemination of scientifically backed ideas.

Keywords: mathematical modeling, epidemiological model, seiz model, sir model, covid-19, twitter, social network analysis, social contagion

Procedia PDF Downloads 38

Authors: Roberto Cabezas H

Abstract:

The current work shows the methodology developed to create narrative lighting spaces for the multimedia performance piece 'cluster: the vanished paradise.' This empirical research is focused on exploring unconventional roles for machines in subjective creative processes, by delving into the semantics of data and machine intelligence algorithms in hybrid technological, creative contexts to expand epistemic domains trough human-machine cooperation. The creative process in scenic and performing arts is guided mostly by intuition; from that idea, we developed an approach to embed collective intuition in computational creative systems, by joining the properties of Generative Adversarial Networks (GAN’s) and Fuzzy Clustering based on a semi-supervised data creation and analysis pipeline. The model makes use of GAN’s to learn from phenomenological data (data generated from experience with lighting scenography) and algorithmic design data (augmented data by procedural design methods), fuzzy logic clustering is then applied to artificially created data from GAN’s to define narrative transitions built on membership index; this process allowed for the creation of simple and complex spaces with expressive capabilities based on position and light intensity as the parameters to guide the narrative. Hybridization comes not only from the human-machine symbiosis but also on the integration of different techniques for the implementation of the aided design system. Machine intelligence tools as proposed in this work are well suited to redefine collaborative creation by learning to express and expand a conglomerate of ideas and a wide range of opinions for the creation of sensory experiences. We found in GAN’s and Fuzzy Logic an ideal tool to develop new computational models based on interaction, learning, emotion and imagination to expand the traditional algorithmic model of computation.

Keywords: fuzzy clustering, generative adversarial networks, human-machine cooperation, hybrid collective data, multimedia performance

Procedia PDF Downloads 122

Authors: P. S. Jagadeesh Kumar, Mingmin Pan, Yang Yung, Tracy Lin Huan

Abstract:

Glaucoma is a group of visual maladies represented by the scheduled optic nerve neuropathy; means to the increasing dwindling in vision ground, resulting in loss of sight. In this paper, a novel support vector machine based retinal therapeutic for glaucoma using machine learning algorithm is conservative. The algorithm has fitting pragmatism; subsequently sustained on correlation clustering mode, it visualizes perfect computations in the multi-dimensional space. Support vector clustering turns out to be comparable to the scale-space advance that investigates the cluster organization by means of a kernel density estimation of the likelihood distribution, where cluster midpoints are idiosyncratic by the neighborhood maxima of the concreteness. The predicted planning has 91% attainment rate on data set deterrent on a consolidation of 500 realistic images of resolute and glaucoma retina; therefore, the computational benefit of depending on the cluster overlapping system pedestal on machine learning algorithm has complete performance in glaucoma therapeutic.

Keywords: machine learning algorithm, correlation clustering mode, cluster overlapping system, glaucoma, kernel density estimation, retinal therapeutic

Procedia PDF Downloads 221

Authors: Shahida Khatoon, Mohd. Faisal Jalil, Vaishali Gautam

Abstract:

The Photovoltaic (PV) generated power is mainly dependent on environmental factors. The PV array’s lifetime and overall systems effectiveness reduce due to the partial shading condition. Clustering the electrical connections between solar modules is a viable strategy for minimizing these power losses by shade diffusion. This article comprehensively evaluates various PV array clustering/reconfiguration models for PV systems. These are static and dynamic reconfiguration techniques for extracting maximum power in mismatch conditions. This paper explores and analyzes current breakthroughs in solar PV performance improvement strategies that merit further investigation. Altogether, researchers and academicians working in the field of dedicated solar power generation will benefit from this research.

Keywords: static reconfiguration, dynamic reconfiguration, photo voltaic array, partial shading, CTC configuration

Procedia PDF Downloads 88

Authors: M. Nekri, A. Khelladi

Abstract:

The world wide web network is a network with a complex topology, the main properties of which are the distribution of degrees in power law, A low clustering coefficient and a weak average distance. Modeling the web as a graph allows locating the information in little time and consequently offering a help in the construction of the research engine. Here, we present a model based on the already existing probabilistic graphs with all the aforesaid characteristics. This work will consist in studying the web in order to know its structuring thus it will enable us to modelize it more easily and propose a possible algorithm for its exploration.

Keywords: clustering coefficient, preferential attachment, small world, web community

Procedia PDF Downloads 246

Authors: Rebecca Cervantes, Elishah Maro Pangilinan

Abstract:

The current research used Uses and gratifications theory to further understand the motivations and satisfaction students get from Facebook and Twitter. The researchers relate the objectives in developing uses and gratifications theory 1) to explain how individuals use mass communication to gratify their needs, “what do people do with the media” many of these young adults use social media networks to communicate with family, friends, and even strangers. Social media sites have created new and non-personal ways for people to interact with others and young adults have taken advantage of this technological trend; 2) to discover underlying motives for individuals’ media use 3) to identify the positive and the negative consequences of individual media use. The researchers use survey questionnaires to gather information that is used in this study. A descriptive analysis was used to measure the answers to a 24-item questionnaire.

Keywords: adolescent, communication, social media, #Hashtag

Procedia PDF Downloads 270

Authors: Chaohong Pan

Abstract:

Through public diplomacy on social media, governments have attempted to influence foreign public opinion. What is the impact of digital public diplomacy? Public diplomacy research often relies on content analysis to study the strategies employed by communicators but has rarely examined its actual impact on the audience. In addition, we do not know if giving a communicator an explicit label, as Twitter does with “government account”, would change the effects of the messages. Can the government label reduce the percussiveness of public diplomacy messages by sending a warning signal? Using a 2 × 2 survey experiment, the present paper contributes to the study of public diplomacy by randomly exposing American participants to four types of tweets from Chinese diplomats. The stimulus materials vary in terms of the tweets’ content (“positive-china” vs. “negative-US) and Twitter government labels (with vs. without the labels). I found that positive tweets about China have a significant positive effect on Americans’ attitudes toward China, whereas negative tweets about the US have little effect on their opinions. Furthermore, positive-China tweets are effective only on China-related issues, which indicates that Chinese diplomats’ tweets have limited effects on shaping a foreign audience’s attitudes toward their own country. Lastly, I find that labels largely have no impact on a diplomatic tweet’s effect. These results contribute to our understanding of the effects of public diplomacy in the digital age.

Keywords: public diplomacy, china, foreign public opinion, twitter

Procedia PDF Downloads 162

Authors: Rikson Gultom

Abstract:

Hate Speech and Abusive language on social media is difficult to detect, usually, it is detected after it becomes viral in cyberspace, of course, it is too late for prevention. An early detection system that has a fairly good accuracy is needed so that it can reduce conflicts that occur in society caused by postings on social media that attack individuals, groups, and governments in Indonesia. The purpose of this study is to find an early detection model on Twitter social media using machine learning that has high accuracy from several machine learning methods studied. In this study, the support vector machine (SVM), Naïve Bayes (NB), and Random Forest Decision Tree (RFDT) methods were compared with the Support Vector machine with genetic algorithm (SVM-GA), Nave Bayes with genetic algorithm (NB-GA), and Random Forest Decision Tree with Genetic Algorithm (RFDT-GA). The study produced a comparison table for the accuracy of the hate speech and abusive language detection model, and presented it in the form of a graph of the accuracy of the six algorithms developed based on the Indonesian-language Twitter dataset, and concluded the best model with the highest accuracy.

Keywords: abusive language, hate speech, machine learning, optimization, social media

Procedia PDF Downloads 107

Authors: Maryam Esfandiari, Bohdan Fridrich

Abstract:

On December 28, 2017, a wave of protests erupted in several Iranian cities. Protesters demonstrated against the president, Hasan Rohani, and theocratical nature of the regime. Iran has a recent history with protest movements, such as Green Movement responsible for demonstrations after 2009 Iranian presidential election. However, the 2017/2018 protests differ from the previous ones in terms of organization and agenda. The events show little to no central organization and seem as being sparked by grass root movements and by citizens’ fatigue of government corruption, authoritarianism, and economic problems of the country. Social media has played important role in communicating the protests to the outside world and also in general coordination. By using content analyses, this paper analyzes the visual content of Twitter posts published during the protests. It aims to find the correlation between their decentralized nature and nature of the tweets – either emotionally arousing or efficiency-elicit. Pictures are searched by hashtags and coded by their content, such as ‘crowds,’ ‘protest activities,’ ‘symbols of unity,’ ‘violence,’ ‘iconic figures,’ etc. The study determines what type of content prevails and what type is the most impactful in terms of reach. This study contributes to understanding the role of social media both as a tool and a space in protest organization and portrayal in countries with limited Internet access.

Keywords: twitter, Iran, collective action, protest

Procedia PDF Downloads 129

Authors: Bahia Braktia, Ana Marcela Montenegro Sanchez

Abstract:

Social media, such as Facebook, Twitter, and YouTube, has been used for teaching and learning for quite some time. These platforms have been proven to be a good tool to improve various language skills, students’ performance of the English language, motivation as well as trigger the authentic language interaction. However, little is known about the potential effects of social media usage on the learning performance of Arabic language learners. The present study explores the potential role that the social media technologies play in learning Arabic as a foreign language at a university in Southeast of United States. In order to investigate this issue, an online survey was administered to examine the perceptions and attitudes of American students learning Arabic. The research questions were: How does social media, specifically Facebook and Twitter, impact the students' Arabic language skills, and what is their attitude toward it? The preliminary findings of the study showed that students had a positive attitude toward the use of social media to enhance their Arabic language skills, and that they used a range of social media features to expose themselves to the Arabic language and communicate in Arabic with native Arabic speaking friends. More detailed findings will be shared in the light data analysis with the audience during the presentation.

Keywords: foreign language learning, social media, students’ perceptions, survey

Procedia PDF Downloads 195

Authors: Thibaud Berthomier, Ali Mansour, Luc Bressollette, Frédéric Le Roy, Dominique Mottier, Léo Fréchier, Barthélémy Hermenault

Abstract:

Deep Venous Thrombosis (DVT) occurs when a thrombus is formed within a deep vein (most often in the legs). This disease can be deadly if a part or the whole thrombus reaches the lung and causes a Pulmonary Embolism (PE). This disorder, often asymptomatic, has multifactorial causes: immobilization, surgery, pregnancy, age, cancers, and genetic variations. Our project aims to relate the thrombus epidemiology (origins, patient predispositions, PE) to its structure using ultrasound images. Ultrasonography and elastography were collected using Toshiba Aplio 500 at Brest Hospital. This manuscript compares two classification approaches: spectral clustering and scattering operator. The former is based on the graph and matrix theories while the latter cascades wavelet convolutions with nonlinear modulus and averaging operators.

Keywords: deep venous thrombosis, ultrasonography, elastography, scattering operator, wavelet, spectral clustering

Procedia PDF Downloads 458

Authors: Kwangmin Joo

Abstract:

Automatic detection of lumbar vertebrae and classification method is proposed for evaluating the degree of disc narrowing. Prior to classification, deep learning based segmentation is applied to detect individual lumbar vertebra. M-net is applied to segment five lumbar vertebrae and fine-tuning segmentation is employed to improve the accuracy of segmentation. Using the features extracted from previous step, clustering technique, k-means clustering, is applied to estimate the degree of disc space narrowing under four grade scoring system. As preliminary study, techniques proposed in this research could help building an automatic scoring system to diagnose the severity of disc narrowing from X-ray images.

Keywords: Disc space narrowing, Degenerative disc disorders, Deep learning based segmentation, Clustering technique

Procedia PDF Downloads 109

Authors: Hae-Yeoun Lee

Abstract:

Quantification of cardiac function is performed by calculating blood volume and ejection fraction in routine clinical practice. However, these works have been performed by manual contouring,which requires computational costs and varies on the observer. In this paper, an automatic left ventricle segmentation algorithm on cardiac magnetic resonance images (MRI) is presented. Using knowledge on cardiac MRI, a K-mean clustering technique is applied to segment blood region on a coil-sensitivity corrected image. Then, a graph searching technique is used to correct segmentation errors from coil distortion and noises. Finally, blood volume and ejection fraction are calculated. Using cardiac MRI from 15 subjects, the presented algorithm is tested and compared with manual contouring by experts to show outstanding performance.

Keywords: cardiac MRI, graph searching, left ventricle segmentation, K-means clustering

Procedia PDF Downloads 383

Authors: H. Aoudia, Y. Touati, E. H. Teguig, A. Ali Cherif

Abstract:

Conceiving and developing routing protocols for wireless sensor networks requires considerations on constraints such as network lifetime and energy consumption. In this paper, we propose a hybrid hierarchical routing protocol named HHRP combining both clustering mechanism and multipath optimization taking into account residual energy and RSSI measures. HHRP consists of classifying dynamically nodes into clusters where coordinators nodes with extra privileges are able to manipulate messages, aggregate data and ensure transmission between nodes according to TDMA and CDMA schedules. The reconfiguration of the network is carried out dynamically based on a threshold value which is associated with the number of nodes belonging to the smallest cluster. To show the effectiveness of the proposed approach HHRP, a comparative study with LEACH protocol is illustrated in simulations.

Keywords: routing protocol, optimization, clustering, WSN

Procedia PDF Downloads 441

Authors: Cherry Yieng Siang Ling, Joong Hee Lee, Myung Hwan Yun

Abstract:

The quality of a product to be usable has become the basic requirement in consumer’s perspective while failing the requirement ends up the customer from not using the product. Identifying usability issues from analyzing quantitative and qualitative data collected from usability testing and evaluation activities aids in the process of product design, yet the lack of studies and researches regarding analysis methodologies in qualitative text data of usability field inhibits the potential of these data for more useful applications. While the possibility of analyzing qualitative text data found with the rapid development of data analysis studies such as natural language processing field in understanding human language in computer, and machine learning field in providing predictive model and clustering tool. Therefore, this research aims to study the application capability of text processing algorithm in analysis of qualitative text data collected from usability activities. This research utilized datasets collected from LG neckband headset usability experiment in which the datasets consist of headset survey text data, subject’s data and product physical data. In the analysis procedure, which integrated with the text-processing algorithm, the process includes training of comments onto vector space, labeling them with the subject and product physical feature data, and clustering to validate the result of comment vector clustering. The result shows 'volume and music control button' as the usability feature that matches best with the cluster of comment vectors where centroid comments of a cluster emphasized more on button positions, while centroid comments of the other cluster emphasized more on button interface issues. When volume and music control buttons are designed separately, the participant experienced less confusion, and thus, the comments mentioned only about the buttons' positions. While in the situation where the volume and music control buttons are designed as a single button, the participants experienced interface issues regarding the buttons such as operating methods of functions and confusion of functions' buttons. The relevance of the cluster centroid comments with the extracted feature explained the capability of text processing algorithms in analyzing qualitative text data from usability testing and evaluations.

Keywords: usability, qualitative data, text-processing algorithm, natural language processing

Procedia PDF Downloads 265

Authors: Josiel Delgadillo, Johnson Kinyua, Charles Mutigwe

Abstract:

The purpose of sentiment analysis is to determine the sentiment strength (e.g., positive, negative, neutral) from a textual source for good decision-making. Natural language processing in domains such as financial markets requires knowledge of domain ontology, and pre-trained language models, such as BERT, have made significant breakthroughs in various NLP tasks by training on large-scale un-labeled generic corpora such as Wikipedia. However, sentiment analysis is a strong domain-dependent task. The rapid growth of social media has given users a platform to share their experiences and views about products, services, and processes, including financial markets. StockTwits and Twitter are social networks that allow the public to express their sentiments in real time. Hence, leveraging the success of unsupervised pre-training and a large amount of financial text available on social media platforms could potentially benefit a wide range of financial applications. This work is focused on sentiment analysis using social media text on platforms such as StockTwits and Twitter. To meet this need, SkyBERT, a domain-specific language model pre-trained and fine-tuned on financial corpora, has been developed. The results show that SkyBERT outperforms current state-of-the-art models in financial sentiment analysis. Extensive experimental results demonstrate the effectiveness and robustness of SkyBERT.

Keywords: BERT, financial markets, Twitter, sentiment analysis

Procedia PDF Downloads 131

Authors: Paul Shize Li, Frank Alber

Abstract:

A fundamental task of clinical genomics is to unravel the functions of genes and their associations with disorders. Although experimental biology has made efforts to discover and elucidate the molecular mechanisms of individual genes in the past decades, still about 40% of human genes have unknown functions, not to mention the diseases they may be related to. For those biologists who are interested in a particular gene with unknown functions, a powerful computational method tailored for inferring the functions and disease relevance of uncharacterized genes is strongly needed. Studies have shown that genes strongly linked to each other in multiple biological networks are more likely to have similar functions. This indicates that the densely connected subgraphs in multiple biological networks are useful in the functional and phenotypic annotation of uncharacterized genes. Therefore, in this work, we have developed an integrative network approach to identify the frequent local clusters, which are defined as those densely connected subgraphs that frequently occur in multiple biological networks and consist of the query gene that has few or no disease or function annotations. This is a local clustering algorithm that models multiple biological networks sharing the same gene set as a three-dimensional matrix, the so-called tensor, and employs the tensor-based optimization method to efficiently find the frequent local clusters. Specifically, massive public gene expression data sets that comprehensively cover dynamic, physiological, and environmental conditions are used to generate hundreds of gene co-expression networks. By integrating these gene co-expression networks, for a given uncharacterized gene that is of biologist’s interest, the proposed method can be applied to identify the frequent local clusters that consist of this uncharacterized gene. Finally, those frequent local clusters are used for function and disease annotation of this uncharacterized gene. This local tensor clustering algorithm outperformed the competing tensor-based algorithm in both module discovery and running time. We also demonstrated the use of the proposed method on real data of hundreds of gene co-expression data and showed that it can comprehensively characterize the query gene. Therefore, this study provides a new tool for annotating the uncharacterized genes and has great potential to assist clinical genomic diagnostics.

Keywords: local tensor clustering, query gene, gene co-expression network, gene annotation

Procedia PDF Downloads 114

Authors: Babak Bahri, Fatemeh Yassaee Meybodi, Changiz Eslahchi

Abstract:

In the pursuit of effective cancer therapies, the exploration of combinatorial drug regimens is crucial to leverage synergistic interactions between drugs, thereby improving treatment efficacy and overcoming drug resistance. However, identifying synergistic drug pairs poses challenges due to the vast combinatorial space and limitations of experimental approaches. This study introduces ClusterSyn, a machine learning (ML)-powered framework for classifying anti-cancer drug synergy scores. ClusterSyn employs a two-step approach involving drug clustering and synergy score prediction using a fully connected deep neural network. For each cell line in the training dataset, a drug graph is constructed, with nodes representing drugs and edge weights denoting synergy scores between drug pairs. Drugs are clustered using the Markov clustering (MCL) algorithm, and vectors representing the similarity of drug pairs to each cluster are input into the deep neural network for synergy score prediction (synergy or antagonism). Clustering results demonstrate effective grouping of drugs based on synergy scores, aligning similar synergy profiles. Subsequently, neural network predictions and synergy scores of the two drugs on others within their clusters are used to predict the synergy score of the considered drug pair. This approach facilitates comparative analysis with clustering and regression-based methods, revealing the superior performance of ClusterSyn over state-of-the-art methods like DeepSynergy and DeepDDS on diverse datasets such as Oniel and Almanac. The results highlight the remarkable potential of ClusterSyn as a versatile tool for predicting anti-cancer drug synergy scores.

Keywords: drug synergy, clustering, prediction, machine learning., deep learning

Procedia PDF Downloads 52

Authors: Didarul Islam Manik

Abstract:

The study was conducted to determine social media usage by the Millennial/young-adult generation in the U.S. and Bangladesh. It investigated what types of social media Millennials/young-adults use in their everyday lives; for what purpose they use social media; what are the significant differences between the two cultures in terms of social media use; and how the age of the respondents correlates with differences in social media use. Among the 409 respondents, 200 were selected from the University of South Dakota and 209 from the University of Dhaka, Bangladesh. The convenience sampling method was used to select the samples. A four-page questionnaire instrument was constructed with 19 closed-ended questions that collected 87 data points. The study considered the uses and gratifications and domestication of technology models as theoretical frameworks. The study found that the Millennials spend an average of 4.5 hours on the Internet daily. They spend an average of 134 minutes on social media every day. However, the U.S. Millennials spend more time (141 minutes) on social media than the Bangladeshis (127 minutes). The U.S. Millennials use various types of social media including Facebook, Twitter, YouTube, Instagram, Pinterest, SnapChat, Reddit, Imgur, etc. In contrast, Bangladeshis use Facebook, YouTube, and Google plus+. The Bangladeshis tended to spend more time on Facebook (107 minutes) than the Americans (57 minutes). The study found that the Millennials of the two countries use Facebook to fill their free time, acquire information, seek entertainment, and maintain existing relationships. However, Bangladeshis are more likely to use Facebook for the acquisition of information, entertainment, educational purposes, and connecting with the people closest to them. Millennials also use Twitter to fill their free time, acquire information, and for entertainment. The study found a statistically significant difference between female and male social media use. It also found a significant correlation between age and using Facebook for educational purposes; age and discussing and posting religious issues; and age and meeting with new people. There is also a correlation between age and the use of Twitter for spending time and seeking entertainment.

Keywords: American study, social media, millennial generation, South Asian studies

Procedia PDF Downloads 210

Authors: Yvonne Dedzo

Abstract:

Social media has become a powerful tool in bridging the distance between individuals regardless of their location. It has become a convenient platform for public discussion and, consequently, generated the phenomenon of citizen journalists who have become both proactive and reactive participants in the dissemination of news, information and other epochal and historical events. This phenomenon has fueled the growth of niche authorities who deliver exceptional democratically consequential information online. This study, therefore, investigates how some selected niche authorities maintain their status on social media. Using the selective processes theory, the study further interrogates the information shared by niche authorities and further analyses the extent to which a public interest-altruistic motive or personal interest-self-serving motive drives their agenda of new sharing and usage. Through cyber-ethnography and, qualitative content analysis and semi-structured interviews, data was gathered and analysed from the posts of two purposely selected niche authorities on Facebook and Twitter. The findings indicate that niche authorities maintain their status by being consistent, prompt, informative, resourceful and interactive in their postings on the social media platform. The study also discovered that even though niche authorities are motivated by both public interest-altruism and interest-self-serving, the latter had a higher of motivation than the former.

Keywords: social medida, citizen journalist, niche authorities, selective processes theory

Procedia PDF Downloads 43

Authors: Qingsong Xiong, Cheng Yuan, Qingzhao Kong, Haibei Xiong

Abstract:

Deep-learning-driven approaches based on vibration responses have attracted larger attention in rapid structural condition assessment while obtaining sufficient measured training data with corresponding labels is relevantly costly and even inaccessible in practical engineering. This study proposes an integrated label propagation network for structural condition assessment, which is able to diffuse the labels from continuously-generating measurements by intact structure to those of missing labels of damage scenarios. The integrated network is embedded with damage-sensitive features extraction by deep autoencoder and pseudo-labels propagation by optimized fuzzy clustering, the architecture and mechanism which are elaborated. With a sophisticated network design and specified strategies for improving performance, the present network achieves to extends the superiority of self-supervised representation learning, unsupervised fuzzy clustering and supervised classification algorithms into an integration aiming at assessing damage conditions. Both numerical simulations and full-scale laboratory shaking table tests of a two-story building structure were conducted to validate its capability of detecting post-earthquake damage. The identifying accuracy of a present network was 0.95 in numerical validations and an average 0.86 in laboratory case studies, respectively. It should be noted that the whole training procedure of all involved models in the network stringently doesn’t rely upon any labeled data of damage scenarios but only several samples of intact structure, which indicates a significant superiority in model adaptability and feasible applicability in practice.

Keywords: autoencoder, condition assessment, fuzzy clustering, label propagation

Procedia PDF Downloads 78

Authors: I Dewa Gede Arya Putra

Abstract:

Principal Component Analysis (PCA) is a mathematical procedure that uses orthogonal transformation techniques to change a set of data with components that may be related become components that are not related to each other. This can have an impact on clustering wind speed characteristics in Indonesia. This study uses data daily wind speed observations of the Site Meteorological Station network for 30 years. Multicollinearity tests were also performed on all of these data before doing clustering with PCA. The results show that the four main components have a total diversity of above 80% which will be used for clusters. Division of clusters using Ward's method obtained 3 types of clusters. Cluster 1 covers the central part of Sumatra Island, northern Kalimantan, northern Sulawesi, and northern Maluku with the climatological pattern of wind speed that does not have an annual cycle and a weak speed throughout the year with a low-speed ranging from 0 to 1,5 m/s². Cluster 2 covers the northern part of Sumatra Island, South Sulawesi, Bali, northern Papua with the climatological pattern conditions of wind speed that have annual cycle variations with low speeds ranging from 1 to 3 m/s². Cluster 3 covers the eastern part of Java Island, the Southeast Nusa Islands, and the southern Maluku Islands with the climatological pattern of wind speed conditions that have annual cycle variations with high speeds ranging from 1 to 4.5 m/s².

Keywords: PCA, cluster, Ward's method, wind speed

Procedia PDF Downloads 174