Latent factor analysis models are an effective type of machine learning model for addressing high-dimensional and sparse matrices, which are encountered in many big-data-related industrial applications. The performance of a latent factor analysis model relies heavily on appropriate hyper-parameters. However, most hyper-parameters are data-dependent, and using grid-search to tune these hyper-parameters is truly laborious and expensive in computational terms. Hence, how to achieve efficient hyper-parameter adaptation for latent factor analysis models has become a significant question.This is the first book to focus on how particle swarm optimization can be incorporated into latent factor analysis for efficient hyper-parameter adaptation, an approach that offers high scalability in real-world industrial applications.
The book will help students, researchers and engineers fully understand the basic methodologies of hyper-parameter adaptation via particle swarm optimization in latent factor analysis models. Further, it will enable them to conduct extensive research and experiments on the real-world applications of the content discussed.
Author(s): Ye Yuan, Xin Luo
Series: SpringerBriefs in Computer Science
Publisher: Springer
Year: 2022
Language: English
Pages: 98
City: Singapore
Preface
Contents
Chapter 1: Introduction
1.1 Background
1.2 Preliminaries
1.2.1 Symbol and Abbreviation Appointment
1.2.2 An LFA Model
1.2.3 Particle Swarm Optimization
1.3 Book Organization
References
Chapter 2: Learning Rate-Free Latent Factor Analysis via PSO
2.1 Overview
2.2 An LFA Model with SGD Algorithm
2.3 The Proposed L2FA Model
2.3.1 Learning Rate Adaptation via PSO
2.3.2 Algorithm Design and Analysis
2.4 Experimental Results and Analysis
2.4.1 General Settings
2.4.2 Performance Comparison
2.4.3 Effect of Swarm Size
2.4.4 Summary
2.5 Conclusions
References
Chapter 3: Learning Rate and Regularization Coefficient-Free Latent Factor Analysis via PSO
3.1 Overview
3.2 An SGD-Based LFA Model
3.3 The Proposed LRLFA Model
3.3.1 Learning Rate and Regularization Coefficient Adaptation via PSO
3.3.2 Linearly Decreasing Inertia Weight Incorporation
3.3.3 Algorithm Design and Analysis
3.4 Experimental Results and Analysis
3.4.1 General Settings
3.4.2 Effect of Swarm Size
3.4.3 Performance Comparison
3.4.4 Summary
3.5 Conclusions
References
Chapter 4: Regularization and Momentum Coefficient-Free Non-negative Latent Factor Analysis via PSO
4.1 Overview
4.2 An SLF-NMU-Based NLFA Model
4.3 The Proposed GALFA Model
4.3.1 A Generalized Learning Objective
4.3.2 A Generalized SLF-NMU-Based Learning Rule
4.3.3 Generalized-Momentum Incorporation
4.3.4 Regularization and Momentum Coefficient Adaptation via PSO
4.3.5 Algorithm Design and Analysis
4.4 Experimental Results and Analysis
4.4.1 General Settings
4.4.2 Parameter Sensitivity
4.4.3 Comparison with State-of-the-Art Models
4.4.4 Summary
4.5 Conclusions
References
Chapter 5: Advanced Learning Rate-Free Latent Factor Analysis via P2SO
5.1 Overview
5.2 An SGD-Based LFA Model
5.3 The Proposed AL2FA Model
5.3.1 A P2SO Algorithm
5.3.2 Learning Rate Adaptation via P2SO
5.3.3 Algorithm Design and Analysis
5.4 Experimental Results and Analysis
5.4.1 General Settings
5.4.2 Effect of ρ
5.4.3 Comparison Results
5.4.4 Summary
5.5 Conclusions
References
Chapter 6: Conclusion and Future Directions
6.1 Conclusion
6.2 Discussion
References
