SOCIAL MEDIA ANALYTICS FOR USER BEHAVIOR MODELING : a task heterogeneity perspective.

Format:

Book

Author/Creator:

Nelakurthi, Arun Reddy.

Contributor:

He, Jingrui.

Series:

Data-enabled engineering.

Data-enabled engineering

Language:

English

Subjects (All):

Machine learning.

Physical Description:

1 online resource (1114 pages).

Edition:

1st edition

Place of Publication:

Boca Raton : CRC Press, 2020.

Summary:

In recent years social media has gained significant popularity and has become an essential medium of communication. Such user-generated content provides an excellent scenario for applying the metaphor of mining any information. Transfer learning is a research problem in machine learning that focuses on leveraging the knowledge gained while solving one problem and applying it to a different, but related problem. Features: Offers novel frameworks to study user behavior and for addressing and explaining task heterogeneity Presents a detailed study of existing research Provides convergence and complexity analysis of the frameworks Includes algorithms to implement the proposed research work Covers extensive empirical analysis Social Media Analytics for User Behavior Modeling: A Task Heterogeneity Perspective is a guide to user behavior modeling in heterogeneous settings and is of great use to the machine learning community.

Contents:

Cover

Half Title

Series Page

Title Page

Copyright Page

Contents

Preface

Acknowledgment

Authors

Contributors

Chapter 1: Introduction

Chapter 2: Literature Survey

2.1 Impact of Social Media

2.2 Heterogeneous Learning and Social Media

2.2.1 Transductive Transfer Learning

2.2.2 Source-free Transfer Learning

2.2.3 Identifying Similar Actors Across Networks

2.3 Explaining Task Heterogeneity

Chapter 3: Social Media for Diabetes Management

3.1 Methodology

3.2 Results

3.3 Discussion

3.4 Challenges in Real-World Applications

Chapter 4: Learning from Task Heterogeneity

4.1 Cross-Domain User Behavior Modeling

4.1.1 Proposed Approach

4.1.1.1 Notation

4.1.1.2 User-Example-Feature Tripartite Graph

4.1.1.3 Objective Function

4.1.1.4 User Soft-Score Weights

4.1.1.5 U-Cross Algorithm

4.1.2 Case Study

4.1.3 Results

4.1.3.1 Data Sets

4.1.3.2 User Selection

4.1.3.3 Empirical Analysis

4.2 Similar Actor Recommendation

4.2.1 Problem Definition

4.2.1.1 Notation and Problem Definition

4.2.2 Proposed Approach

4.2.2.1 Matrix Factorization for Cross Network Link Recommendation

4.2.2.2 Proposed Framework

4.2.2.3 Optimization Algorithm

4.2.2.4 Link Recommendation

4.2.2.5 Complexity Analysis

4.2.3 Results

4.2.3.1 Data Sets

4.2.3.2 Experiment Setup

4.2.3.3 Case Study

4.3 Source-Free Domain Adaptation

4.3.1 Problem Definition

4.3.2 Proposed Approach

4.3.2.1 Label Deficiency

4.3.2.2 Distribution Shift

4.3.2.3 Convergence of AOT

4.3.3 Results

4.3.3.1 Two Stage Analysis

4.3.3.2 Sensitivity Analysis

4.3.3.3 Convergence Analysis

4.3.3.4 Runtime Analysis

Chapter 5: Explainable Transfer Learning

5.1 Proposed Approach

5.1.1 Notation

5.1.2 exTL Framework

5.1.3 Reweighting the Source Domain Examples.

5.1.4 Domain Invariant Representation

5.1.5 Algorithm

5.1.6 Shallow Neural Network: An Example

5.2 Results

5.2.1 Text Data

5.2.2 Images

Chapter 6: Conclusion

6.1 User Behavior Modeling in Social Media

6.2 Addressing and Explaining Task Heterogeneity

6.3 Limitations

6.3.1 Addressing Concept Drift

6.3.2 Model Fairness

6.3.3 Negative Transfer

6.3.4 Ethical Issues in Healthcare

6.3.5 Misinformation and Disinformation in Healthcare

6.4 Future Work

Bibliography

Index.

Notes:

OCLC-licensed vendor bibliographic record.

ISBN:

1-000-02536-5

0-429-27035-6

1-000-02540-3

9780429270352

OCLC:

1137835525