Review Comment Analysis For E-commerce

by Rong Zhang, Aoying Zhou, Wenzhe Yu, Yifan Gao, Pingfu Chao

Overview

This book presents the recent achievements on the processing of representative user generated content (UGC) on E-commerce websites. This large size of UGC is valuable information for data mining to help customer/object profiling. It provides a comprehensive overview on the concept of customer credibility, object-oriented review summarization technology and content-based collaborative filtering algorithm. It covers a feedback mechanism which is designed to discover customer credibility, which is used to define the professional degree of review content; product-oriented review summarization for restaurants or trip arrangements, and introduced content-based collaborative filtering for product recommendation.

Product Details

ISBN-13:	9789813100046
Publisher:	World Scientific Publishing Company, Incorporated
Publication date:	08/29/2016
Series:	East China Normal University Scientific Reports , #5
Pages:	172
Product dimensions:	6.00(w) x 9.10(h) x 0.70(d)

Table of Contents

Preface ix

Acknowledgments xiii

1 Introduction 1

1.1 Background 1

1.2 Challenges 2

1.3 Related Work 9

1.4 Outline of Book Content 17

2 Credibility Learning 19

2.1 Problem Definition 19

2.1.1 Problem Description 20

2.1.2 Background 21

2.2 Scoring Framework Overview 24

2.3 Review Comment Analysis 25

2.3.1 ME Model 26

2.3.2 Constructing Labeled Data 27

2.3.3 Training and Prediction 29

2.4 Customer Credibility Calculation 29

2.4.1 Product and Customer 30

2.4.2 Shop and Customer 32

2.4.3 Credibility Calculation 33

2.5 Re-scoring 33

2.6 Experimental Results 34

2.6.1 Datasets 34

2.6.2 Review Comment Analysis 35

2.6.3 Product and Shop Re-scoring 38

2.7 Conclusion 44

3 Entity Resolution 45

3.1 Problem Definition 45

3.2 Learning-based Method on Centralized System 47

3.2.1 Data Preprocessing 48

3.2.2 Entity Resolution 54

3.3 Random-based Method on Distributed System 56

3.3.1 Framework Introduction 57

3.3.2 Entity Signature Generation 58

3.3.3 Candidate Pair Generation 59

3.3.4 Redundancy Reduction 61

3.4 Experimental Results 63

3.4.1 Results of Learning-based Method 63

3.4.2 Results of Random-based Method 67

3.5 Conclusion 74

4 Review Selection 75

4.1 Problem Definition 75

4.2 Quality and Diversity of Review Set 76

4.2.1 The Quality of Review Set 76

4.2.2 The Diversity of Review Set 78

4.3 Review Selection Algorithm 80

4.3.1 Opinion Extraction 80

4.3.2 Review Selection 81

4.4 Experimental Results 81

4.4.1 Dataset and Settings 81

4.4.2 Evaluation on Diversity Factor 83

4.4.3 Comparison of Algorithms 86

4.5 Conclusion 88

5 Review Summarization 89

5.1 Problem Definition 8^9

5.1.1 Diversify-based Summary Generation 90

5.1.2 Topic Model-based Summary Generation 91

5.2 Diversity-based Approach 92

5.2.1 Finding Evaluative Snippets 92

5.2.2 Predicting Snippet Scores 94

5.2.3 Summarizing Product Snippets 95

5.3 Topic Model-based Approach 99

5.3.1 Bilateral Topic Model 100

5.3.2 Inference Algorithm 104

5.3.3 Summarization 108

5.4 Experimental Results 109

5.4.1 Dataset 110

5.4.2 Results of Diversity-based Approach 111

5.43 Results of Topic Model-based Approach 117

5.5 Conclusion 124

6 Recommendation 127

6.1 Background 127

6.1.1 Traditional Methods 127

6.1.2 Topic Analysis 128

6.2 Comment-based Collaborative Filtering Model 129

6.2.1 Profile Generator 130

6.2.2 Representation of Samples 131

6.2.3 Prediction Models 133

6.2.4 Enhanced Systems 134

6.2.5 Representative Review Selection 134

6.3 Experimental Results 135

6.3.1 Dataset 136

6.3.2 Evaluation Metric 136

6.3.3 Baselines 137

6.3.4 Topic Analysis 137

6.3.5 Main Results 138

6.3.6 Further Analysis 139

6.4 Conclusion 139

7 Conclusion 141

Bibliography 143

Index 153