Machine Learning Algorithms and Applications

Mettu Srinivas PhD from the Indian Institute of Technology Hyderabad, and is currently an assistant professor in the Department of Computer Science and Engineering, NIT Warangal, India.G. Sucharitha PhD from KL University, Vijayawada and is currently an assistant professor in the Department of Electronics and Communication Engineering at ICFAI Foundation for Higher Education Hyderabad. Anjanna Matta PhD from the Indian Institute of Technology Hyderabad and is currently an assistant professor in the Department of Mathematics at ICFAI Foundation for Higher Education Hyderabad. Prasenjit Chatterjee PhD is an associate professor in the Mechanical Engineering Department at MCKV Institute of Engineering, India.

• Acknowledgments xvPreface xviiPart 1: Machine Learning for Industrial Applications 11 A Learning-Based Visualization Application for Air Quality Evaluation During COVID-19 Pandemic in Open Data Centric Services 3Priyank Jain and Gagandeep Kaur1.1 Introduction 41.1.1 Open Government Data Initiative 41.1.2 Air Quality 41.1.3 Impact of Lockdown on Air Quality 51.2 Literature Survey 51.3 Implementation Details 61.3.1 Proposed Methodology 71.3.2 System Specifications 81.3.3 Algorithms 81.3.4 Control Flow 101.4 Results and Discussions 111.5 Conclusion 21References 212 Automatic Counting and Classification of Silkworm Eggs Using Deep Learning 23Shreedhar Rangappa, Ajay A. and G. S. Rajanna2.1 Introduction 232.2 Conventional Silkworm Egg Detection Approaches 242.3 Proposed Method 252.3.1 Model Architecture 26.3.2 Foreground-Background Segmentation 282.3.3 Egg Location Predictor 302.3.4 Predicting Egg Class 312.4 Dataset Generation 352.5 Results 352.6 Conclusion 37Acknowledgment 38References 383 A Wind Speed Prediction System Using Deep Neural Networks 41Jaseena K. U. and Binsu C. Kovoor3.1 Introduction 423.2 Methodology 453.2.1 Deep Neural Networks 453.2.2 The Proposed Method 473.2.2.1 Data Acquisition 473.2.2.2 Data Pre-Processing 483.2.2.3 Model Selection and Training 503.2.2.4 Performance Evaluation 513.2.2.5 Visualization 513.3 Results and Discussions 523.3.1 Selection of Parameters 523.3.2 Comparison of Models 533.4 Conclusion 57References 574 Res-SE-Net: Boosting Performance of ResNets by Enhancing Bridge Connections 61Varshaneya V., S. Balasubramanian and Darshan Gera4.1 Introduction 614.2 Related Work 624.3 Preliminaries 634.3.1 ResNet 634.3.2 Squeeze-and-Excitation Block 644.4 Proposed Model 664.4.1 Effect of Bridge Connections in ResNet 664.4.2 Res-SE-Net: Proposed Architecture 674.5 Experiments 684.5.1 Datasets 684.5.2 Experimental Setup 684.6 Results 694.7 Conclusion 73References 745 Hitting the Success Notes of Deep Learning 77Sakshi Aggarwal, Navjot Singh and K.K. Mishra5.1 Genesis 785.2 The Big Picture: Artificial Neural Network 795.3 Delineating the Cornerstones 805.3.1 Artificial Neural Network vs. Machine Learning 805.3.2 Machine Learning vs. Deep Learning 815.3.3 Artificial Neural Network vs. Deep Learning 815.4 Deep Learning Architectures 825.4.1 Unsupervised Pre-Trained Networks 825.4.2 Convolutional Neural Networks 835.4.3 Recurrent Neural Networks 845.4.4 Recursive Neural Network 855.5 Why is CNN Preferred for Computer Vision Applications? 855.5.1 Convolutional Layer 865.5.2 Nonlinear Layer 865.5.3 Pooling Layer 875.5.4 Fully Connected Layer 875.6 Unravel Deep Learning in Medical Diagnostic Systems 895.7 Challenges and Future Expectations 945.8 Conclusion 94References 956 Two-Stage Credit Scoring Model Based on Evolutionary Feature Selection and Ensemble Neural Networks 99Diwakar Tripathi, Damodar Reddy Edla, Annushree Bablani and Venkatanareshbabu Kuppili6.1 Introduction 1006.1.1 Motivation 1006.2 Literature Survey 1016.3 Proposed Model for Credit Scoring 1036.3.1 Stage-1: Feature Selection 1046.3.2 Proposed Criteria Function 1056.3.3 Stage-2: Ensemble Classifier 1066.4 Results and Discussion 1076.4.1 Experimental Datasets and Performance Measures 1076.4.2 Classification Results With Feature Selection 1086.5 Conclusion 112References 1137 Enhanced Block-Based Feature Agglomeration Clustering for Video Summarization 117Sreeja M. U. and Binsu C. Kovoor7.1 Introduction 1187.2 Related Works 1197.3 Feature Agglomeration Clustering 1227.4 Proposed Methodology 1227.4.1 Pre-Processing 1237.4.2 Modified Block Clustering Using Feature Agglomeration Technique 1257.4.3 Post-Processing and Summary Generation 1277.5 Results and Analysis 1297.5.1 Experimental Setup and Data Sets Used 1297.5.2 Evaluation Metrics 1307.5.3 Evaluation 1317.6 Conclusion 138References 138Part 2: Machine Learning for Healthcare Systems 1418 Cardiac Arrhythmia Detection and Classification From ECG Signals Using XGBoost Classifier 143Saroj Kumar Pandeyz, Rekh Ram Janghel and Vaibhav Gupta8.1 Introduction 1438.2 Materials and Methods 1458.2.1 MIT-BIH Arrhythmia Database 1468.2.2 Signal Pre-Processing 1478.2.3 Feature Extraction 1478.2.4 Classification 1488.2.4.1 XGBoost Classifier 1488.2.4.2 AdaBoost Classifier 1498.3 Results and Discussion 1498.4 Conclusion 155References 1569 GSA-Based Approach for Gene Selection from Microarray Gene Expression Data 159Pintu Kumar Ram and Pratyay Kuila9.1 Introduction 1599.2 Related Works 1619.3 An Overview of Gravitational Search Algorithm 1629.4 Proposed Model 1639.4.1 Pre-Processing 1639.4.2 Proposed GSA-Based Feature Selection 1649.5 Simulation Results 1669.5.1 Biological Analysis 1689.6 Conclusion 172References 172Part 3: Machine Learning for Security Systems 17510 On Fusion of NIR and VW Information for Cross-Spectral Iris Matching 177Ritesh Vyas, Tirupathiraju Kanumuri, Gyanendra Sheoran and Pawan Dubey10.1 Introduction 17710.1.1 Related Works 17810.2 Preliminary Details 17910.2.1 Fusion 18110.3 Experiments and Results 18210.3.1 Databases 18210.3.2 Experimental Results 18210.3.2.1 Same Spectral Matchings 18310.3.2.2 Cross Spectral Matchings 18410.3.3 Feature-Level Fusion 18610.3.4 Score-Level Fusion 18910.4 Conclusions 190References 19011 Fake Social Media Profile Detection 193Umita Deepak Joshi, Vanshika, Ajay Pratap Singh, Tushar Rajesh Pahuja, Smita Naval and Gaurav Singal11.1 Introduction 19411.2 Related Work 19511.3 Methodology 19711.3.1 Dataset 19711.3.2 Pre-Processing 19811.3.3 Artificial Neural Network 19911.3.4 Random Forest 20211.3.5 Extreme Gradient Boost 20211.3.6 Long Short-Term Memory 20411.4 Experimental Results 20411.5 Conclusion and Future Work 207Acknowledgment 207References 20712 Extraction of the Features of Fingerprints Using Conventional Methods and Convolutional Neural Networks 211E. M. V. Naga Karthik and Madan Gopal12.1 Introduction 21212.2 Related Work 21312.3 Methods and Materials 21512.3.1 Feature Extraction Using SURF 21512.3.2 Feature Extraction Using Conventional Methods 21612.3.2.1 Local Orientation Estimation 21612.3.2.2 Singular Region Detection 21812.3.3 Proposed CNN Architecture 21912.3.4 Dataset 22112.3.5 Computational Environment 22112.4 Results 22212.4.1 Feature Extraction and Visualization 22312.5 Conclusion 226Acknowledgements 226References 22613 Facial Expression Recognition Using Fusion of Deep Learning and Multiple Features 229M. Srinivas, Sanjeev Saurav, Akshay Nayak and Murukessan A. P.13.1 Introduction 23013.2 Related Work 23213.3 Proposed Method 23513.3.1 Convolutional Neural Network 23613.3.1.1 Convolution Layer 23613.3.1.2 Pooling Layer 23713.3.1.3 ReLU Layer 23813.3.1.4 Fully Connected Layer 23813.3.2 Histogram of Gradient 23913.3.3 Facial Landmark Detection 24013.3.4 Support Vector Machine 24113.3.5 Model Merging and Learning 24213.4 Experimental Results 24213.4.1 Datasets 24213.5 Conclusion 245Acknowledgement 245References 245Part 4: Machine Learning for Classification and Information Retrieval Systems 24714 AnimNet: An Animal Classification Network using Deep Learning 249Kanak Manjari, Kriti Singhal, Madhushi Verma and Gaurav Singal14.1 Introduction 24914.1.1 Feature Extraction 25014.1.2 Artificial Neural Network 25014.1.3 Transfer Learning 25114.2 Related Work 25214.3 Proposed Methodology 25414.3.1 Dataset Preparation 25414.3.2 Training the Model 25414.4 Results 25814.4.1 Using Pre-Trained Networks 25914.4.2 Using AnimNet 25914.4.3 Test Analysis 26014.5 Conclusion 263References 26415 A Hybrid Approach for Feature Extraction From Reviews to Perform Sentiment Analysis 267Alok Kumar and Renu Jain15.1 Introduction 26815.2 Related Work 26915.3 The Proposed System 27115.3.1 Feedback Collector 27215.3.2 Feedback Pre-Processor 27215.3.3 Feature Selector 27215.3.4 Feature Validator 27415.3.4.1 Removal of Terms From Tentative List of Features on the Basis of Syntactic Knowledge 27415.3.4.2 Removal of Least Significant Terms on the Basis of Contextual Knowledge 27615.3.4.3 Removal of Less Significant Terms on the Basis of Association With Sentiment Words 27715.3.4.4 Removal of Terms Having Similar Sense 27815.3.4.5 Removal of Terms Having Same Root 27915.3.4.6 Identification of Multi-Term Features 27915.3.4.7 Identification of Less Frequent Feature 27915.3.5 Feature Concluder 28115.4 Result Analysis 28215.5 Conclusion 286References 28616 Spark-Enhanced Deep Neural Network Framework for Medical Phrase Embedding 289Amol P. Bhopale and Ashish Tiwari16.1 Introduction 29016.2 Related Work 29116.3 Proposed Approach 29216.3.1 Phrase Extraction 29216.3.2 Corpus Annotation 29416.3.3 Phrase Embedding 29416.4 Experimental Setup 29716.4.1 Dataset Preparation 29716.4.2 Parameter Setting 29716.5 Results 29816.5.1 Phrase Extraction 29816.5.2 Phrase Embedding 29816.6 Conclusion 303References 30317 Image Anonymization Using Deep Convolutional Generative Adversarial Network 305Ashish Undirwade and Sujit Das17.1 Introduction 30617.2 Background Information 31017.2.1 Black Box and White Box Attacks 31017.2.2 Model Inversion Attack 31117.2.3 Differential Privacy 31217.2.3.1 Definition 31217.2.4 Generative Adversarial Network 31317.2.5 Earth-Mover (EM) Distance/Wasserstein Metric 31617.2.6 Wasserstein GAN 31717.2.7 Improved Wasserstein GAN (WGAN-GP) 31717.2.8 KL Divergence and JS Divergence 31817.2.9 DCGAN 31917.3 Image Anonymization to Prevent Model Inversion Attack 31917.3.1 Algorithm 32117.3.2 Training 32217.3.3 Noise Amplifier 32317.3.4 Dataset 32417.3.5 Model Architecture 32417.3.6 Working 32517.3.7 Privacy Gain 32517.4 Results and Analysis 32617.5 Conclusion 328References 329Index 331