Deep Learning-Based Approaches for Sentiment Analysis (Algorithms for Intelligent Systems)

Preface
Contents
About the Editors
Application of Deep Learning Approaches for Sentiment Analysis
	1 Introduction
	2 Taxonomy of Sentiment Analysis
		2.1 Sentiment Analysis, Polarity, and Output
		2.2 Levels of Sentiment Analysis
		2.3 Domain Applicability, Training, and Testing Strategy
		2.4 Language Support
		2.5 Evaluation Measures
	3 Text Representation for Sentiment Analysis
		3.1 Embedded Vectors
		3.2 Strategy of Initializing the Embedded Vectors
		3.3 Enhancing the Embedded Vectors
		3.4 Approximation Methods
		3.5 Sampling-Based Approaches
		3.6 Softmax-Based Approaches
	4 Deep Learning Approaches for Sentiment Analysis
	5 Evaluation Metrics for Sentiment Analysis
	6 Benchmarked Datasets and Tools
	7 Conclusion
	References
Recent Trends and Advances in Deep Learning-Based Sentiment Analysis
	1 Introduction
	2 Related Work
	3 Machine Learning Approaches for Sentiment Analysis
	4 Study Rationale
	5 Deep Learning Architectures
		5.1 Convolutional Neural Networks
		5.2 Recurrent Neural Networks
		5.3 Bi-directional Recurrent Neural Network
	6 Long Short-Term Memory (LSTMs)
	7 Gated Recurrent Units (GRUs)
	8 Attention Mechanism
	9 Research Methodology
	10 Approach to Sentiment Analysis Task Categorization
	11 Coarse-Grain Sentiment Analysis
	12 Fine-Grain Sentiment Analysis
	13 Cross-Domain Sentiment Analysis
	14 Conclusion and Survey Highlights
	References
Deep Learning Adaptation with Word Embeddings for Sentiment Analysis on Online Course Reviews
	1 Introduction
	2 State of the Art
		2.1 Sentiment Analysis in E-Learning Systems
		2.2 Deep Learning for Sentiment Analysis
		2.3 Word Embeddings for Sentiment Analysis
	3 Word Embedding Representations for Text Mining
		3.1 Word2Vec
		3.2 GloVe
		3.3 FastText
		3.4 Intel
	4 Deep Learning Components for Text Mining
		4.1 Feed-Forward Neural Network (FNN)
		4.2 Recurrent Neural Network (RNN)
		4.3 Long Short-Term Memory (LSTM) Network
		4.4 Convolutional Neural Network (CNN)
		4.5 Normalization Layer (NL)
		4.6 Attention Layer (AL)
		4.7 Other Layers
	5 Our Sentiment Predictor for E-Learning Reviews
		5.1 Review Splitting
		5.2 Word Embedding Modeling
		5.3 Review Vectorization
		5.4 Sentiment Model Definition
		5.5 Sentiment Model Training and Prediction
	6 Experimental Evaluation
		6.1 Dataset
		6.2 Baselines
		6.3 Metrics
		6.4 Deep Neural Network Model Regressor Performance
		6.5 Contextual Word Embeddings Performance
	7 Conclusions, Open Challenges, and Future Directions
	References
Toxic Comment Detection in Online Discussions
	1 Online Discussions and Toxic Comments
		1.1 News Platforms and Other Online Discussions Forums
		1.2 Classes of Toxicity
	2 Deep Learning for Toxic Comment Classification
		2.1 Comment Datasets for Supervised Learning
		2.2 Neural Network Architectures
	3 From Binary to Fine-Grained Classification
		3.1 Why Is It a Hard Problem?
		3.2 Transfer Learning
		3.3 Explanations
	4 Real-World Applications
		4.1 Semi-automated Comment Moderation
		4.2 Troll Detection
	5 Current Limitations and Future Trends
		5.1 Misclassification of Comments
		5.2 Research Directions
	6 Conclusions
	References
Aspect-Based Sentiment Analysis of Financial Headlines and Microblogs
	1 Introduction
	2 Related Work
	3 State-of-the-Art Models
		3.1 ALA Model
		3.2 IIIT Delhi Model
	4 Our Methodology
		4.1 Features
	5 Aspect Classification Models
		5.1 Models
		5.2 Classification Model Training
	6 Sentiment Models
		6.1 Models
		6.2 Sentiment Model Training
	7 Evaluation
		7.1 Data Set
		7.2 Data Augmentation
		7.3 Data Pre-processing
		7.4 Metrics
		7.5 Results
	8 Conclusion and Future Work
	References
Deep Learning-Based Frameworks for Aspect-Based Sentiment Analysis
	1 Introduction
	2 Problem Formulation
		2.1 Aspect-Term Extraction
		2.2 Aspect-Category Detection
	3 Observation/Assumption in ABSA
	4 Input Representation
	5 Concepts Related to Deep Learning
		5.1 Word-Embeddings
		5.2 Long Short-Term Memory (LSTM)
		5.3 Bi-directional Long Short-Term Memory (Bi-LSTM)
		5.4 RNN with Attention
		5.5 Convolution Neutral Network (CNN)
	6 Deep Learning Architectures Used in ABSA
		6.1 Sentiment Analysis
		6.2 Aspect-Term Extraction
		6.3 Aspect-Category Extraction
		6.4 Aspect-Based Sentiment Detection
	7 Conclusion
	References
Transfer Learning for Detecting Hateful Sentiments in Code Switched Language
	1 Introduction
		1.1 Hate Speech Problem
		1.2 Code Switched and Code Mixed Languages
		1.3 Challenges in Code Switched and Code Mixed Languages
		1.4 Deep Learning
		1.5 Overview
	2 Background and Related Work
		2.1 Language Identification
		2.2 POS Tagging
		2.3 Named Entity Recognition
		2.4 Sentiment Analysis
	3 Dataset and Evaluation
		3.1 HOT Dataset
		3.2 Bohra et al. bohra2018dataset dataset
		3.3 HEOT Dataset
		3.4 Davidson Dataset
	4 Methodology
		4.1 SVM and Random Forest
		4.2 Ternary Trans-CNN Model
		4.3 LSTM-Based Model
		4.4 MIMCT Model
	5 Results
		5.1 SVM and Random Forest Classifier
		5.2 Ternary Trans-CNN Model
		5.3 LSTM Model with Transfer Learning
		5.4 MIMCT Model
	6 Conclusion
	7 Future Work
	References
Multilingual Sentiment Analysis
	1 Introduction
		1.1 Low Resource Language
		1.2 Challenges of Sentiment Analysis
		1.3 Deep Learning
	2 Literature Survey
		2.1 High Resource Languages
		2.2 Lexicon-Based Approaches
		2.3 Traditional Machine Learning-Based Approaches
		2.4 Low Resource Languages
	3 Word Embeddings for Sentiment Analysis
		3.1 Refining Word Embeddings for Sentiment Analysis
		3.2 Improving Word Embedding Coverage in Low Resource Languages
	4 Deep Learning Techniques for Multilingual Sentiment Analysis
		4.1 Convolutional Neural Networks
		4.2 Recurrent Neural Networks
		4.3 Autoencoders
		4.4 Bilingual Constrained Recursive Autoencoders
		4.5 AROMA
		4.6 Siamese Neural Networks
	5 Discussion
	6 Conclusion
	References
Sarcasm Detection Using Deep Learning-Based Techniques
	1 Introduction
	2 Related Work
	3 Grice’s Maxims
	4 Challenges in Sarcasm Detection
	5 Dataset Description
	6 Feature Description
	7 Process Outline
	8 Models Used
	9 Experiments and Results
	10 Future Scope
	References
Deep Learning Approaches for Speech Emotion Recognition
	1 Introduction
	2 Feature Extraction
	3 Feature Selection
	4 Classical Approaches
		4.1 Speaker-Dependent SER
		4.2 Speaker-Independent SER
		4.3 Other Models
	5 Deep Learning Approaches
	6 System Overview
		6.1 Classical Approach for SER
		6.2 Deep Learning Approaches for SER
		6.3 Critical Comparision
	7 Evaluation
		7.1 Dataset Description
		7.2 Original Results
		7.3 Results Obtained
	8 Comparison of Existing Approaches
	9 Conclusions
	References
Bidirectional Long Short-Term Memory-Based Spatio-Temporal in Community Question Answering
	1 Introduction
	2 Related Works
	3 Methodology
		3.1 Preprocessing Steps
		3.2 Best Answer Prediction
	4 Experimental Setup: Answer Classification
	5 Experiment II: Answer Ranking
	6 Conclusion
	References
Comparing Deep Neural Networks to Traditional Models for Sentiment Analysis in Turkish Language
	1 Introduction
	2 Methodology
	3 Experimental Setup and Results
		3.1 Dataset
		3.2 Traditional BOW Approach
		3.3 Deep Learning Architecture
	4 Conclusion
	References