Big Data, Machine Learning, and Applications: Proceedings of the 2nd International Conference, BigDML 2021 (Lecture Notes in Electrical Engineering, 1053)

✍ Scribed by Malaya Dutta Borah; Dolendro Singh Laiphrakpam; Nitin Auluck; Valentina Emilia Balas

• Tongue: English
• Leaves: 758
• Category: Library

✦ Table of Contents

Contents
About the Editors
Android Application-Based Security Surveillance Implementing Machine Learning
1 Introduction
2 Proposed system
3 Components Required
3.1 Hardware Requirements
3.2 Software Requirements
4 Hardware Implementation
5 Working of the System
6 Software Implementation
7 Results and Discussion
8 Conclusion
References
Realtime Object Distance Measurement Using Stereo Vision Image Processing
1 Introduction
2 Related Work
3 Proposed Methodology
3.1 Pre-processing of an Image
3.2 Object Detection and Segmentation
3.3 Object Distance and Size Measurement Using Stereo Camera
3.4 Mathematical Substantiation for Object Detection and Localization
4 Result and Discussion
5 Conclusion
References
An Insight on Drone Applications in Surveillance Domain
1 Introduction
2 Literature Survey
2.1 Related Work
2.2 Problem Definition
2.3 Key Challenges
3 The Model
3.1 Architecture/Model
3.2 Components/Elements—Protocols
3.3 Working Principle
3.4 Applications
4 Discussions
5 Conclusion and Future Scope
References
Handwritten Mixed Numerals Classification System
1 Introduction
2 Review of Literature
3 Proposed Approach
4 Result Analysis
5 Conclusion
References
IoT Based Smart Farm Monitoring System
1 Introduction
2 Literature Review
3 Proposed System
3.1 An App Connected to the IoT Based System Installed in the Farm
3.2 Automatic Gate Closing if Left Open
3.3 Smart Irrigation
3.4 Automatic Greenhouse Monitoring and Control System Project
3.5 Tracking the Crops’ Condition by Detecting Data from the Calendar, and the Weather of the Place Using Application Programming Interface (API’s) and Suggest to the Farmer the Kind of Care to be Taken
4 Methodology
5 Experimental Setup
6 Conclusion
References
An Extensive Review of the Supervised Learning Algorithms for Spiking Neural Networks
1 Introduction
2 Review of Supervised Learning Methods
2.1 Learning by Finding the Gradient
2.2 Asymmetric Supervised Hebbian Learning
2.3 Learning with Remote Supervision
2.4 Learning with Metaheuristics
3 Conclusion
References
Multitask Learning-Based Simultaneous Facial Gender and Age Recognition with a Weighted Loss Function
1 Introduction
2 Literature Review
3 Proposed Method
4 Implementation Details
5 Results
6 Conclusion
References
Visualizing Crime Hotspots by Analysing Online Newspaper Articles
1 Introduction
2 Related Work
3 Definition of Terms
3.1 Named Entity Recognition
3.2 BERT
3.3 BIO Scheme
4 Proposed System
4.1 Web Scraping
4.2 Dataset
4.3 Location and Crime Extraction
5 Experimental Results
5.1 Evaluation of Model
5.2 Visualization of Results
6 Conclusion and Future Scope
References
Applications of Machine Learning for Face Mask Detection During COVID-19 Pandemic
1 Introduction
2 Literature Review
3 Research Background on Developed Algorithm
3.1 VGGNet
3.2 ResNet
3.3 InceptionNet
4 Proposed Neural Network and Its Architectural Framework
4.1 Proposed System
4.2 Data Acquisition
4.3 Testing and Training Datasets
4.4 Encoding Approaches
4.5 Decoding Approaches
5 Statistical Techniques and Software Adopted
5.1 Statistical Tools
5.2 Software Used
6 Results
6.1 Data Analysis
6.2 Findings
6.3 Confusion Matrix Outcomes
6.4 Evalution Metrics
7 Conclusion
References
A Cascaded Deep Learning Approach for Detection and Localization of Crop-Weeds in RGB Images
1 Introduction
2 Literature Review
3 Proposed Work
3.1 Data Description and Pre-processing
3.2 Weeds Classification Step
3.3 Weeds Segmentation Step
4 Results and Analysis
4.1 Experimental Setup
4.2 Classification
4.3 Semantic Segmentation
5 Conclusion
References
Ensemble of Deep Learning Enabled Tamil Handwritten Character Recognition Model
1 Introduction
2 Related Works
3 The Proposed EDL-THCR Technique
3.1 Preprocessing
3.2 Ensemble of DL Models
3.3 Softmax Layer
4 Performance Validation
5 Conclusion
References
A Comparative Study of Loss Functions for Deep Neural Networks in Time Series Analysis
1 Introduction
2 Related Work
3 Deep Neural Network and Its Architecture
3.1 Convolutional Neural Networks
3.2 Recurrent Neural Networks and Their Variants
4 Loss Functions of DNNs in Regression
5 Experiments
5.1 Experimental Setup
5.2 Datasets Selection
5.3 Selected Models and Loss Functions
5.4 Performance Evaluation
6 Comparative Analysis of Loss Functions in Time Series
7 Conclusion
References
Learning Algorithm for Threshold Softmax Layer to Handle Unknown Class Problem
1 Introduction
2 Related Work
3 Threshold Softmax Layer
3.1 Softmax Function
3.2 Threshold Learning Algorithm
3.3 Handling Unknown Class Problem Using TSM Layer
4 Experiment and Results
4.1 Experimental Setup
4.2 Experimental Results
5 Conclusion
References
Traffic Monitoring and Violation Detection Using Deep Learning
1 Introduction
2 Proposed Method
2.1 YOLO
2.2 SORT
2.3 Vehicle Counting
2.4 Violations
3 Conclusion
4 Future Scope
References
Conjugate Gradient Method for finding Optimal Parameters in Linear Regression
1 Introduction
2 Related Works
3 Linear Regression
4 Results and Discussion
4.1 Boston Housing Dataset
4.2 Selecting Features From the Dataset
4.3 Result Analysis
5 Conclusion
References
Rugby Ball Detection, Tracking and Future Trajectory Prediction Algorithm
1 Introduction
2 Methodology
2.1 Section I
2.2 Section II
2.3 Section III
3 Results
4 Conclusion
5 Future Scope
References
Early Detection of Heart Disease Using Feature Selection and Classification Techniques
1 Introduction
2 Literature Review
3 Materials and Methods
3.1 Dataset
3.2 Methodology
4 Results and Discussion
5 Conclusion
References
Gun Detection System for Surveillance Cameras Using HOG-Assisted KNN Classifier
1 Introduction
2 State Of The Art
2.1 Challenges
3 Proposed Model
4 Experimental Results and Analysis
5 Conclusions and Future Work
References
Optimized Detection, Classification, and Tracking with YOLOV5, HSV Color Thresholding, and KCF Tracking
1 Introduction
2 Methodology 
2.1 Section I
2.2 Section II
2.3 Section III
3 Results
4 Conclusion
5 Future Scope
References
COVID-19 Detection Using Chest X-ray Images
1 Introduction
2 Related Works
3 Methodology
3.1 Dataset Preparation 
3.2 Preprocessing
3.3 Feature Extraction
3.4 Classifier
3.5 Optimised Feature Selection
4 Results and Discussions
5 Conclusion
References
Comparative Analysis of LDA Algorithm for Low Resource Indian Languages with Its Translated English Documents
1 Introduction
2 Literature Review
2.1 Importance of Indian Native Languages
2.2 Twitter
2.3 Topic Modeling
2.4 Latent Dirichlet Allocation (LDA)
3 Methodology
3.1 Data Collection
3.2 Data Cleaning
3.3 Data Pre-processing
3.4 Parameter Selection
3.5 Experiments Using LDA
3.6 Evaluation
4 Experiments
4.1 Data Collection and Data Preprocessing
4.2 Model Hyperparameters Selection
4.3 Evaluation
5 Conclusions
References
Text Style Transfer: A Comprehensive Study on Methodologies and Evaluation
1 Introduction
2 Datasets
2.1 Parallel and Non-parallel Data
2.2 Available benchmark Datasets for TST
3 Methodologies
3.1 Parallel
3.2 Non-parallel
3.3 Unsupervised Methods
4 Evaluation Techniques
4.1 Automatic Evaluation
4.2 Human Evaluation
5 Discussion and Conclusion
References
Classification of Hindustani Musical Ragas Using One-Dimensional Convolutional Neural Networks
1 Introduction
2 Related Work
3 Dataset
4 Methodology
4.1 Artificial Neural Network (ANN)
4.2 1D Convolutional Neural Network (1D-CNN)
5 Experimentation and Results
6 Conclusion
References
W-Tree: A Concept Correlation Tree for Data Analysis and Annotations
1 Introduction
2 Literature Survey
3 W-Tree Model
3.1 Design Principles
3.2 Data Set
3.3 Model Design
3.4 Algorithms
4 Results and Discussion
4.1 Scraping the Articles
4.2 Building a Data Dictionary
4.3 Graph Generation
4.4 Aggregating Five Domains
4.5 Annotations
4.6 Overall Analysis
4.7 Time and Space Complexity Analysis
5 Conclusion
References
Crawl Smart: A Domain-Specific Crawler
1 Introduction
2 Literature Survey
3 Crawl Smart Model
3.1 Design Principles
3.2 Model Design
3.3 Data Design
3.4 Abstract Data Type Representation
3.5 Annotations
3.6 Similarity Module
3.7 Knowledge Base
4 Results and Discussion
5 Conclusion and Future Scope
References
Evaluating the Effect of Leading Indicators in Customer Churn Prediction
1 Introduction
2 Literature Survey
2.1 Churn Prediction Using Multiple Data Sources
2.2 Early Churn Detection
3 Experimental Setup
3.1 Dataset
3.2 Methodology
4 Experiments, Results, and Discussion
4.1 Effect of Leading Indicators
4.2 Lead Time Analysis
5 Conclusion and Future Work
References
Classification of Skin Lesion Using Image Processing and ResNet50
1 Introduction
2 Literature Survey
3 Methodology
3.1 Hair Removal
3.2 Image Augmentation
3.3 Segmentation
4 Experimentation and Result
4.1 Accuracy and Loss Graph
4.2 ROC Curve
4.3 Confusion Matrix
5 Conclusion
References
Data Collection and Pre-processing for Machine Learning-Based Student Dropout Prediction
1 Introduction
2 State of the Art
3 Data Sources
3.1 Primary Data
3.2 Secondary Data
4 Approaches and Techniques of Data Collection
4.1 Acquisition of Data
4.2 Labeling of Data
4.3 Improving Existing Data/Model
5 Flowchart for Data Collection
6 Data Pre-processing Steps
6.1 Data Cleaning
6.2 Feature Extraction
6.3 Data Transformation/Normalization
6.4 Feature Selection
7 Conclusions
References
Nested Named-Entity Recognition in Multilingual Code-Switched NLP
1 Introduction
2 Related Work
2.1 Named-Entity Recognition (NER)
2.2 Data Augmentation
3 Dataset
3.1 Adaptive Fine-Tuning
3.2 Behavioral Fine-Tuning
3.3 Task-Specific Fine-Tuning for NER
4 Method
4.1 Pre-trained Language Model
4.2 Data Augmentation
4.3 Adaptive Fine-Tuning
4.4 Behavioral Fine-Tuning
4.5 Task-Specific Fine-Tuning
5 Results
6 Discussions
7 Conclusion
References
Deep Learning-Based Semantic Segmentation of Blood Cells from Microscopic Images
1 Introduction
2 Related Works
3 Dataset
4 Methodology
4.1 Preprocessing
4.2 Semantic Segmentation Framework
5 Experimentation and Results
6 Conclusion
References
A Partitioned Task Offloading Approach for Privacy Preservation at Edge
1 Introduction
2 Related Work
3 Proposed Methodology
3.1 Denoising AutoEncoder
3.2 Task Offloading
4 Results and Discussion
4.1 Data Corruption
4.2 Comparison of Performance
5 Conclusion and Future Work
References
Artificial Intelligence in Radiological COVID-19 Detection: A State-of-the-Art Review
1 Introduction
2 Review Paper Based on Deep Learning Techniques
2.1 Convolutional Neural Network-Based COVID-19 Detection
2.2 Transfer Learning and Image Segmentation CNN-Based COVID-19 Detection
2.3 Traditional Machine Learning-Based COVID-19 Detection
3 Summary of Studied Work
4 Comparative Analysis
5 Conclusion and Future Work
References
Anomaly Detection in SCADA Industrial Control Systems Using Bi-Directional Long Short-Term Memory
1 Introduction
2 Related Work
3 The New Gas Pipeline Dataset
4 Feature Engineering
4.1 Handling Missing Values
4.2 Preprocessing Techniques
4.3 Feature Scaling
4.4 Handling Imbalanced Data
4.5 Parameter Setting for Bi-LSTM/Bi-GRU Model
5 Proposed Methodology
5.1 LSTM (Versus) GRU
5.2 Bi-LSTM
5.3 Bi-GRU
6 Implementation Detail
6.1 Data Preparation
6.2 Evaluation Metrics
7 Result Analysis and Discussion
8 Conclusion and Future Work
References
Implementing Autonomous Navigation on an Omni Wheeled Robot Using 2D LiDAR, Tracking Camera and ROS
1 Introduction
1.1 Introduction
1.2 Robot Operating System
2 Robot Base Description
3 System Architecture
3.1 ROS Navigation Stack
3.2 ROS-Arduino Communication
3.3 Robot Control
3.4 ROS Mobile
4 Experimental Set Up
4.1 Setting Up the Robot Base
4.2 Running the Code
4.3 Providing Destination Co-ordinates for Autonomous Navigation
5 Results
6 Conclusion
References
Analysis of Deep Learning Models for Text Summarization of User Manuals
1 Introduction
2 Related Work
3 System Overview
3.1 Indexer
3.2 Retriever
3.3 Summarizer
4 Performance Analysis
4.1 Dataset
4.2 Dataset Statistics
4.3 Analysis of Extractive Models
4.4 PEGASUS Summary Generation and Load Time Analysis
4.5 Key Observations
4.6 Hit Ratio
4.7 Statistical Analysis
5 Results
5.1 Rouge Score Analysis
6 Conclusion and Future Work
References
Modelling Seismic Performance of Reinforced Concrete Buildings Within Response Spectrum Framework
1 Introduction
2 Framework for Response Spectrum Analysis
3 Numerical Examples
4 Conclusions
References
A Survey on DDoS Detection Using Deep Learning in Software Defined Networking
1 Introduction
2 DDoS Attack on SDN
2.1 SDN Architecture
2.2 Various Deep Learning Algorithms
2.3 Background of DDoS Attack
3 Deep Learning-Based Intrusion Detection Techniques in SDN
4 Analysis
4.1 Based on Methodology
4.2 Based on Evaluation Techniques
4.3 Based on Dataset
5 Conclusion
References
Segmentation of Dentin and Enamel from Panoramic Dental Radiographic Image (OPG) to Detect Tooth Wear
1 Introduction
2 Experimental Method
2.1 Image Pre-processing
2.2 Image Enhancement
2.3 Segmentation
3 Results and Discussion
4 Conclusion
References
Revisiting Facial Key Point Detection—An Efficient Approach Using Deep Neural Networks
1 Introduction
2 Literature Review
3 Research Methodology
3.1 Dataset Description
3.2 Image Pre-processing
3.3 Imputation Techniques
3.4 Data Augmentation
3.5 Inference Time
3.6 Loss Functions
3.7 Evaluation Metrics
3.8 Model Architecture
4 Results
4.1 Evaluation of RMSE Scores
4.2 Evaluation of Model Size and Parameters
4.3 Inference Time Analysis
4.4 Evaluation of Training Curves
4.5 Visualization of Test Images
5 Conclusion
References
A Hybrid Framework Using Natural Language Processing and Collaborative Filtering for Performance Efficient Feedback Mining and Recommendation
1 Introduction
2 Literary Review
3 Methodology and Survey Results
3.1 Survey Design
3.2 Employing Participants
3.3 Survey Participants
3.4 Survey Analysis
4 Proposed Work
4.1 Data Collection and Data Preprocessing
4.2 Domain Classification of Reviews
4.3 Sentiment Analysis
5 Routing Feedback
6 Recommendation Systems
7 Results
8 Conclusion and Future Work
References
Facial Recognition-Based Automatic Attendance Management System Using Deep Learning
1 Introduction
1.1 Review of the Literature
2 Methodology
2.1 Models Used
3 Dataset Details
3.1 Dataset Collection
3.2 Pre-processing
3.3 Training Process
3.4 Dataset Preparation/Used
4 Results and Discussions
5 Conclusion
References
Application of Infrared Thermography in Assessment of Diabetic Foot Anomalies: A Treatise
1 Introduction
2 Plantar Foot Temperature and Thermography
3 Thermal Distribution—Normal Versus Diabetic Foot
4 Thermal Analysis of Foot with Diabetic Peripheral Neuropathy
5 Thermal Analysis of Foot with Peripheral Arterial Disease
6 Factors Affecting Acquisition of Thermal Images—for Plantar Foot Analysis
6.1 Emissivity
6.2 Reflectivity
6.3 Ambient Temperature and Humidity
6.4 Angle of Incidence
6.5 Patient Side Factors
7 Conclusion
References
A Survey and Classification on Recommendation Systems
1 Introduction
1.1 Classification
1.2 Contributions of the Paper
2 Literature Survey
2.1 Content-Based Filtering Method
2.2 Collaborative Filtering Method
2.3 Knowledge-Based Filtering Method
2.4 Hybrid Method
2.5 Emotion-Based Recommendation System
3 Emerging Trends and Future Scope
4 Conclusion
References
Analysis of Synthetic Data Generation Techniques in Diabetes Prediction
1 Introduction
2 Related Works
2.1 Method
2.2 SMOTE Oversampling
2.3 SVM-SMOTE Oversampling
2.4 ADASYN Oversampling
2.5 SMOTETomek Data Generation
2.6 SMOTEENN Synthetic Data Generation
2.7 Proposed Approach of SDG
2.8 Results
3 Discussion
4 Conclusion
References
Beyond Information Exchange: An Approach to Deploy Network Properties for Information Diffusion
1 Introduction
1.1 Basic Idea
1.2 Contributions
2 Method
2.1 Common Neighborhood
2.2 Common Neighborhood Strategy
3 Experimental Analysis
3.1 Result Analysis
4 Conclusion
References
Sentiment Analysis on Worldwide COVID-19 Outbreak
1 Introduction
2 Literature Survey
3 Architecture
4 Methodology
4.1 Data Collection and Analysis
4.2 Data Pre-processing
4.3 World Cloud Analysis
4.4 Sentiment Classification Using Bert
4.5 Evaluation Metrics
5 Results and Analysis
6 Conclusion and Future Work
References
Post-Vaccination Risk Prediction of COVID-19: Machine Learning Approach
1 Introduction
1.1 Motivation
1.2 Related Work
2 Problem Statement
3 Proposed Architecture and Methodology
3.1 Data Collection
3.2 Data-Preprocessing
3.3 Splitting of Dataset
3.4 Classification
3.5 Performance Evaluation Metrics
4 Comparative Results and Discussions
4.1 Support Vector Machine
4.2 Logistic Regression
4.3 Random Forest
5 Conclusion
References
Offensive Language Detection in Under-Resourced Algerian Dialectal Arabic Language
1 Introduction
2 Related Work
3 Corpus
3.1 Data Collection
3.2 Data Annotation
4 Experimental Results
5 Conclusion
References
A Comparative Analysis of Modern Machine Learning Approaches for Automatic Classification of Scientific Articles
1 Introduction
2 Data and Methods
2.1 Dataset Description
2.2 Methodology
2.3 Data Preprocessing
2.4 ML Classifiers
2.5 DL Classifiers
3 Experimental Results and Analysis
3.1 Experimental Setup
4 Conclusion
References
A Review of Machine Learning Algorithms on Different Breast Cancer Datasets
1 Introduction
2 Datasets: WDBC, WBCD, and SEER
2.1 Wisconsin Diagnostics Breast Cancer (WDBC) Dataset
2.2 Wisconsin Breast Cancer Database-Original (WBCD) Dataset
2.3 Surveillance, Epidemiology, and End Results (SEER) Dataset
3 ML Algorithms for Breast Cancer Classification and Survival Rate Prediction
4 Conclusion
Annexure-Expansion of Acronyms Mentioned in Sect.  3
References
52 The Online Behaviour of the Algerian Abusers in Social Media Networks
Abstract
1 Introduction
2 Related Work
3 Algerian Social Media Users
3.1 Online Social Behaviour
3.2 Writing Behaviour
4 Algerian Online Abusers Behaviour
4.1 Statistics
4.2 Abusers Activities on Their Accounts
4.3 Writing Style
5 Conclusion
References
Interactive Attention AI to Translate Low-Light Photos to Captions for Night Scene Understanding in Women Safety
1 Introduction
1.1 Need and Significance
1.2 Contributions
1.3 Novelty
1.4 Research Gap and Related Work
2 AI Architecture, Methods and Results
2.1 AI for Describing Night Scenes for Visually Impaired Users
2.2 Novel Result: This Neural Network Translates Low-Brightness Images to Sentences!
2.3 Dataset
2.4 Novelty in AI Inference for Interactive Caption Generation
2.5 HMI in Attention Model for Interactive Caption Generation
2.6 Feasibility Towards Enabling Women Safety Apps
3 Conclusions
References
AI Visualization in Nanoscale Microscopy
1 Introduction
1.1 The Opportunity: Applications of AI to New Fields
1.2 Research Gap
1.3 Literature Review and Novelty
1.4 Contributions
2 Methods and Results
2.1 Open-Source AI at Enabling Discoveries in the World of Nanoscale
2.2 Seeing a Nanomaterial Through an “AI Lens”
2.3 Visualizing Nanoscale Patterns Learnt by CNN
3 Conclusions and Future Directions
References
Convolutional Gated MLP: Combining Convolutions and gMLP
1 Introduction
1.1 Related Literature and Research Directions
1.2 Research Gap/Novelty
2 Methods and Results
2.1 Contributions
2.2 Convolution Gated MLP
2.3 Comparison of gMLP Versus CgMLP
2.4 Experimental Results: GMLP Versus CgMLP
2.5 How the Convolutional Gated MLP Learns?
2.6 GMLP Versus CgMLP: Spatial Interactions Versus Feature Channel Interactions
2.7 GMLP Versus CgMLP: Inductive Biases
2.8 Source Code Availability and Reproducible Results
3 Summary
References
Unique Covariate Identity (UCI) Detection for Emotion Recognition Through EEG Signals
1 Introduction
1.1 Electroencephalography (EEG)
1.2 Methods
1.3 Feature Extraction
1.4 Machine Learning Approaches
1.5 Linear Discriminant Analysis (LDA)
1.6 Hidden Markov Model (HMM)
1.7 Support Vector Machines (SVM)
2 Literature Survey
3 System Tool
3.1 Datasets Available
4 System Design
4.1 DWT: Discrete Wavelet Transform
4.2 MNR: Multi-nominal Regression
4.3 SOM: Self-organized Mapping Optimizer
5 Results and Discussions
5.1 Preprocessing of EEG Signals
6 Challenges
7 Conclusion
References
A Simple and Effective Method for Segmenting Lung Regions from CT Scan Images Using K-Means
1 Introduction
2 K-Means Clustering Using Recursive Averaging
2.1 Recursive Averaging
2.2 K-Means Clustering Using Recursive Averaging
3 Segmentation of Lung Regions from CT Scan Images
4 Experimental Results
5 Conclusions
References
Risk-Based Portfolio Optimization on Some Selected Sectors of the Indian Stock Market
1 Introduction
2 Related Work
3 The Portfolio Design Methodology
4 Performance Evaluation
4.1 The Media Sector
4.2 The Oil and Gas Sector
4.3 The Private Sector Banks
4.4 The PSU Banks
4.5 The Realty Sector
5 Conclusion
References