Intersection of Machine Learning and Computational Social Sciences

Cover....1 Half Title....2 Series Page....3 Title Page....4 Copyright Page....5 Table of Contents....6 Preface....8 Editors....9 List of Contributors....11 Chapter 1: Leveraging artificial intelligence for educational transformation: A critical study of the Indian context....14 1.1 Introduction....14 1.2 Review of literature....18 1.3 Application of AI in education....25 1.4 Application of AI in Indian education system....27 1.5 Critical analysis of AI in Indian education system....34 1.6 Discussion....37 1.7 Conclusion....42 References....43 Chapter 2: A study on artificial intelligence and its role in medical image analysis....47 2.1 Introduction....47 2.1.1 Deep learning in medical imaging....50 2.1.2 Limitations and challenges....54 2.1.3 Proposed solutions....58 2.1.4 Future directions....61 2.2 Conclusion....61 References....62 Chapter 3: An overview of machine learning: Concepts, algorithms, and applications....68 3.1 Introduction....68 3.2 Motivation....69 3.3 Novel contributions in the chapter....69 3.4 Chapter organization....70 3.5 Categorization of machine learning....70 3.6 Machine learning algorithms....73 3.7 Machine learning application....76 3.8 Ethical considerations in machine learning....78 3.9 Model evaluation and selection....79 3.10 Feature engineering and selection....80 3.11 Case studies and future directions and trends in machine learning....81 3.12 Conclusion....83 References....84 Chapter 4: Natural language processing: Food habits-based disease prediction using large language models....88 4.1 Introduction....88 4.2 Food style and healthcare....88 4.3 Large language models (LLMs)....89 4.4 Proposed methodology....92 4.5 Experimental model development....95 4.6 Result and discussion....99 4.7 Conclusion....106 References....106 Chapter 5: Convolutional neural network-based plant leaf disease classification: Implications for society and agriculture....109 5.1 Introduction....109 5.2 Precision agriculture....110 5.3 Literature review....111 5.4 Datasets....113 5.4.1 PlantVillage dataset....113 5.4.2 Rice leaf dataset....113 5.5 Results and analysis....115 5.5.1 Evaluation metrics....115 5.5.2 Experiments on the PlantVillage dataset....116 5.5.3 Experiments on rice dataset....120 5.6 Conclusions....125 References....125 Chapter 6: Classifying the social world: Algorithms and applications in computational social science....128 6.1 Introduction....128 6.1.1 The challenge of understanding the social world....128 6.1.2 The rise of computational social science....129 6.1.3 The power of classification by unveiling patterns in social data....129 6.2 Applications of classification algorithms in social science....130 6.2.1 Mapping social networks....130 6.2.2 Understanding public opinion through sentiment analysis of textual data....131 6.2.3 Beyond words: image classification for social research....132 6.2.3.1 Facial expression or emotion recognition....132 6.2.3.2 Scene analysis....132 6.2.4 Classifying economic behavior through consumer trends and market analysis....133 6.2.4.1 Consumer trend analysis....133 6.2.4.2 Customer churn prediction....133 6.2.4.3 Market analysis and fraud detection....133 6.2.5 Applications across disciplines....134 6.3 Common classification algorithms in computational social science....134 6.3.1 Supervised learning techniques....134 6.3.1.1 Logistic regression....135 6.3.1.2 Decision trees....135 6.3.1.3 Support vector machines (SVMs)....136 6.3.1.4 Neural networks....137 6.3.2 Unsupervised learning techniques....138 6.3.2.1 K-means clustering....138 6.3.2.2 Hierarchical clustering....139 6.3.3 Choosing the right algorithm....140 6.4 Validation and beyond....141 6.4.1 The need for validation and ensuring model accuracy and generalizability....141 6.4.2 Evaluating classification performance through various metrics....141 6.4.3 Cross-validation techniques: testing models on unseen data....142 6.4.4 Addressing bias and fairness....143 6.5 Case studies in computational social science classification....143 6.5.1 Predicting customer churn in the retail industry....144 6.6 Challenges and ethical considerations in computational social sciences....148 6.6.1 Challenges in computational social sciences....148 6.6.2 Ethical considerations in computational social sciences....149 6.7 Future of computational social science....149 6.8 Conclusion....150 References....151 Chapter 7: Social network analysis: Need, data collection, APIs, data preprocessing, feature engineering techniques, etc.....154 7.1 Introduction....154 7.2 Initial network design....155 7.2.1 Modeling of the network....156 7.2.2 Network characteristics....156 7.2.2.1 Structural characteristics....157 7.2.2.1.1 Density....157 7.2.2.1.2 Size....157 7.2.2.1.3 Diversity....157 7.2.2.1.4 Structural hole....157 7.2.2.1.5 Clique....157 7.2.2.1.6 Degree....157 7.2.2.1.7 Betweenness....157 7.2.2.1.8 Closeness....158 7.2.2.1.9 Clustering coefficients....158 7.2.2.1.10 Isolates....158 7.2.2.1.11 PageRank....158 7.2.2.2 Relational characteristics....158 7.2.2.2.1 Strength of tie....158 7.2.2.2.2 Availability....159 7.2.2.2.3 Appealability....159 7.2.2.2.4 Trust....159 7.2.2.2.5 Reputation....159 7.2.2.2.6 Reliance....159 7.2.2.2.7 Expected mutual aid....159 7.2.2.2.8 Rivalry....160 7.2.2.3 Individual characteristics....160 7.2.2.3.1 Character....160 7.2.2.3.2 Emotional intelligence....160 7.2.2.3.3 Purposefulness....161 7.2.2.3.4 Prior experience....161 7.2.2.3.5 Emotional dissection....161 7.3 SNA process....161 7.3.1 Design....163 7.3.2 Data collection....163 7.3.3 Data analysis and visualization....163 7.3.4 Data review....163 7.3.5 Evaluation....164 7.3.6 Take actions....164 7.3.7 Summary....164 7.4 SNA software and tools....165 7.4.1 SocNetV....166 7.4.2 NetworkX....167 7.4.3 Pajek....167 7.4.4 Cytoscape....167 7.4.5 Apache Spark GraphX....167 7.4.6 JGraphT....167 7.4.7 UCINET....167 7.4.8 igraph....168 7.4.9 JUNG....168 7.5 Using APIs (application programming interfaces) to retrieve data....168 7.5.1 How to retrieve data using APIs....169 7.5.1.1 Choose the appropriate API....169 7.5.1.2 Authentication and access....169 7.5.1.3 Requesting and retrieving data....169 7.5.1.4 Data storage and parsing....169 7.5.1.5 Data management....169 7.6 Feature engineering....170 7.6.1 Customized feature engineering techniques for network data....170 7.6.1.1 Hole structures....170 7.6.1.2 Clustering parameters....170 7.6.1.3 Algorithms for community identification....170 7.6.1.4 Centrality indicators....171 7.6.2 Advanced methods: node embeddings and graph neural networks....171 7.7 Anticipated research findings....172 7.7.1 Information distribution....172 7.7.1.1 The threshold model....172 7.7.1.2 The cascade models....172 7.7.1.3 The epidemiological model....172 7.7.1.4 Model of triggering....173 7.7.1.5 Time-aware model....173 7.7.2 Link prediction....173 7.7.2.1 Predicated on resemblance....173 7.7.2.2 Based on maximum likelihood and probability....173 7.7.2.3 Reduction of dimensionality....174 7.7.3 Influence maximization....174 7.7.3.1 Based on approximations....174 7.7.3.2 Heuristics....174 7.7.3.3 Based in the community....175 7.7.3.4 Heuristics in meta-form....175 7.7.4 Community detection....175 7.7.4.1 Partitioning graphs....176 7.7.4.2 Dynamic approaches....176 7.7.4.3 Grouping in hierarchies....176 7.7.4.4 Methods based on density....176 7.7.5 Big data using SNA....177 7.7.6 Summary....178 7.8 Social network insights platform....178 7.8.1 Education....179 7.8.1.1 Identifying experts in the field....179 7.8.1.2 Examining learner performance....179 7.8.1.3 Evaluation of institutions and women researchers relevance....180 7.8.1.4 Recognizing new interest-based communities across the faculty....180 7.8.2 Sports....180 7.8.2.1 Football....180 7.8.2.2 Cricket....180 7.8.3 Society and culture....181 7.8.3.1 Gathering social network information for writing....181 7.8.3.2 Examining storylines and adapted films....181 7.8.3.3 Investigating media and news....181 7.8.4 Patterns and knowledge discovery....182 7.8.5 Politics....182 7.8.5.1 False information and fake news....182 7.8.5.2 Campaigning for elections....182 7.8.6 Tourism and hospitality....182 7.8.7 Promotion and branding....183 7.8.8 Disaster management....183 7.8.9 Transport....183 7.8.10 Cyber security....184 7.8.11 Multimedia....184 7.9 Challenges and future directions....184 7.9.1 Variability....184 7.9.2 Flexibility....184 7.9.3 Relation type and weight....184 7.9.4 Features of the node....185 7.9.5 Restrictions on data....185 7.9.6 Insufficient public datasets....185 7.9.7 Analysis in context....185 7.9.8 Maintaining privacy....185 7.9.9 Including topological network architecture....185 7.9.10 The spread and evolution of opinions....185 7.10 Conclusion....186 References....186 Chapter 8: Feature selection in DNA microarray data: Insights for healthcare and social science applications through machine learning....195 8.1 Introduction....195 8.2 Microarray data....197 8.2.1 What is a microarray dataset?....197 8.2.2 Essential properties inherent to microarray data....197 8.2.2.1 Small sample size....198 8.2.2.2 Class imbalance....198 8.2.2.3 Data shift....198 8.2.2.4 Outliers....199 8.2.3 Datasets and repositories....199 8.3 Feature selection on microarray data....200 8.3.1 Filter-based algorithms....201 8.3.1.1 ReliefF....201 8.3.1.2 Correlation-based feature selection (CFS)....202 8.3.1.3 Information gain (IG)....202 8.3.1.4 Minimal-redundancy-maximal-relevance (mRMR)....203 8.3.1.5 Rough sets and fuzzy-rough sets-based feature selection....204 8.3.2 Wrapper-based algorithms....205 8.3.3 Embedded algorithms....206 8.3.4 Other algorithms....207 8.4 An empirical setup....208 8.4.1 Selected datasets and feature selection methods....209 8.4.2 Evaluation metrics....209 8.5 Results analysis and discussion....212 8.5.1 Cancer RNA-seq dataset....212 8.5.1.1 Results analysis....212 8.5.1.2 Discussion....214 8.5.2 Leukemia dataset....216 8.5.2.1 Results analysis....216 8.5.2.2 Discussion....217 8.5.3 Ovarian dataset....221 8.5.3.1 Results analysis....221 8.5.3.2 Discussion....223 8.5.4 Limitations and challenges....227 8.6 Conclusion....229 References....230 Chapter 9: Self-supervised learning for pathological speech detection....234 9.1 Introduction....234 9.2 Contextual embeddings....236 9.3 Methodology....237 9.3.1 Dataset....237 9.3.2 Implementation....237 9.3.3 Embedding extraction....238 9.4 Results and discussion....238 9.5 Conclusion....240 References....241 Chapter 10: Analyzing the social consequences of lung cancer risk prediction with lifestyle data: A comparative study of machine learning techniques....244 10.1 Introduction....244 10.1.1 Contribution....245 10.2 Review and related literature....245 10.3 Methodology....246 10.3.1 Parameter information....247 10.3.2 Dataset description....247 10.3.3 Exploratory data analysis....248 10.4 Findings and discussion....251 10.4.1 Confusion matrix....251 10.4.2 Accuracy of the model....254 10.4.3 Other statistical measurements....255 10.4.4 AUC-ROC curve....255 10.4.5 Comparative analysis....258 10.4.6 Social impact....259 10.5 Conclusion and future work....259 Disclosure of correspondence....260 Data accessibility declaration....260 Note....260 References....260 Chapter 11: Adversarial learning for enhancing security in Cobot-driven industries: A machine learning approach to risk mitigation....262 11.1 Introduction to Cobots....262 11.2 Application of Cobots....264 11.2.1 Advantages of adopting Cobots in industries....266 11.3 Security challenges in collaborative robots....268 11.3.1 Examining the hazards of collaborative robot environments....269 11.4 Literature review....271 11.4.1 Types of cooperative situations....272 11.4.2 Examples of real businesses....273 11.4.3 Various applications of HRC....274 11.4.4 Programming in Cobots....276 11.5 Adversarial attacks on Cobots....280 11.5.1 Understanding adversarial attacks....280 11.5.2 Impacts of adversarial attacks on Cobots....280 11.5.3 Adversarial attack mitigation strategies....280 11.6 Adversarial learning for Cobots security....281 11.7 Challenges and limitations....283 11.7.1 Addressing challenges and limitations....284 11.8 Future trends and research directions....284 11.9 Conclusion....287 References....287 Chapter 12: Cybersecurity challenges in energy harvesting systems: A machine learning approach to safeguarding industrial IoT networks....292 12.1 Introduction....292 12.2 Basic concepts....294 12.2.1 Energy harvesting techniques....294 12.2.2 Energy harvesting via solar energy....296 12.2.3 Energy harvesting via radio frequency....297 12.3 Cyberattacks and threats in EH network....298 12.4 Wiretapping....299 12.5 Denial of service (DoS)....299 12.5.1 Side-channel attack....300 12.5.2 Spoofing....301 12.5.3 De-authentication....301 12.6 Physical-layer data secrecy attack....302 12.6.1 Untrusted relays....304 12.6.2 Internal adversaries....304 12.6.3 Lightweight cryptography technique....304 12.7 Literature survey....305 12.8 Current cybersecurity methods in energy harvesting....307 12.8.1 Security methods....307 12.9 Probable attacks on energy harvesting....308 12.10 Future research directions....309 12.11 Conclusion....310 References....310 Chapter 13: Integrating transfer learning techniques for automated recognition of medicinal plant leaves in computational social science....314 13.1 Introduction and importance of medicinal plants....314 13.1.1 Manual plant identification challenges....314 13.1.2 Machine learning and deep learning potential for automated plant identification....314 13.1.3 Contribution of this study....315 13.2 Related work....316 13.3 Proposed design....320 13.4 Dataset description....320 13.5 Data preprocessing....321 13.6 Feature extraction and machine learning approaches....321 13.7 Machine learning classifiers....322 13.7.1 Logistic regression....322 13.7.2 Random Forest....322 13.7.3 Support vector machine....323 13.8 Classification....324 13.8.1 Evaluation criteria....324 13.8.2 Experiment and results....325 13.8.3 Discussion....331 13.8.4 Challenges and proposed solutions....332 13.9 Conclusion....332 References....333 Chapter 14: Deep learning-based approach for combating fake news....336 14.1 Introduction....336 14.2 Literature review....337 14.3 Methodology....338 14.3.1 Data collection....338 14.3.2 Data preprocessing....339 14.3.3 Data augmentation....340 14.3.4 Tokenization....341 14.3.5 Normalization....341 14.3.6 Removal of noise, URLs, hashtags, and user mentions....341 14.3.7 Word segmentation....341 14.3.8 Replacing emoticons and emojis....342 14.3.9 Abbreviations and slang....342 14.3.10 Punctuation removal....343 14.3.11 Stopword removal....343 14.3.12 Stemming and lemmatization....343 14.4 Feature engineering....344 14.4.1 TFIDF....344 14.4.2 Bag of words....345 14.4.3 Sentiment analysis....345 14.4.4 News length....345 14.5 Pseudocode 1 EFEFI: hybrid feature engineering approach for FakeNews identification....346 14.6 Deep learning methods used....347 14.6.1 Convolutional neural networks (CNNs)....347 14.6.2 Long short-term memory (LSTM) networks....347 14.7 Results and analysis....348 14.7.1 Accuracy....348 14.7.2 Precision....349 14.7.3 Recall....349 14.7.4 F1-score....350 14.7.4.1 LSTM....350 14.7.4.2 CNN....350 14.7.4.3 Ensemble (CNN and LSTM)....350 14.8 Conclusion....357 References....357 Index....360