This book dives into the inner workings of recommender systems, those ubiquitous technologies that shape our online experiences. From Netflix show suggestions to personalized product recommendations on Amazon or the endless stream of curated YouTube videos, these systems power the choices we see every day.
Collaborative filtering reigns supreme as the dominant approach behind recommender systems. This book offers a comprehensive exploration of this topic, starting with memory-based techniques. These methods, known for their ease of understanding and implementation, provide a solid foundation for understanding collaborative filtering. As you progress, you'll delve into latent factor models, the abstract and mathematical engines driving modern recommender systems.
The journey continues with exploring the concepts of metadata and diversity. You'll discover how metadata, the additional information gathered by the system, can be harnessed to refine recommendations. Additionally, the book delves into techniques for promoting diversity, ensuring a well-balanced selection of recommendations. Finally, the book concludes with a discussion of cutting-edge deep learning models used in recommender systems.
This book caters to a dual audience. Firstly, it serves as a primer for practicing IT professionals or data scientists eager to explore the realm of recommender systems. The book assumes a basic understanding of linear algebra and optimization but requires no prior knowledge of machine learning or programming. This makes it an accessible read for those seeking to enter this exciting field.Secondly, the book can be used as a textbook for a graduate-level course. To facilitate this, the final chapter provides instructors with a potential course plan.
Foreword
Preface
Chapter 1: Introduction and Organization
Introduction
Content of this book
Chapter 2: Neighborhood Based Models
Introduction
User-based Approach
Item-based Approach
References
Chapter 3: Ratings
Introduction
Biases and Baseline Correction
Significance Weighting
Optimally Learnt Interpolation Weights
References
Chapter 4: Latent Factor Models
Introduction
Latent Factor Model
Nuclear Norm Minimization
References
Chapter 5: Using Metadata
Introduction
Matrix Factorization on Graphs
Nuclear norm minimization on Multiple Graphs
Label consistent nuclear norm minimization
Label consistent matrix factorization
References
Appendix
Chapter 6: Diversity in Recommender Systems
Introduction
Prior Art
Matrix Factorization based Diversity Model
Nuclear Norm-based Diversity Model
References
Chapter 7: Deep Latent Factor Models
Introduction
Brief Introduction to Representation Learning
Deep Latent Factor Model
Graphical Deep Latent Factor Model
Diversity in Deep Latent Factor Model
References
Chapter 8: Conclusion & Note to Instructors
Introduction
Course Organization
Expectation from pupils
Evaluation
Biography
Angshul’s research interests lie in signal processing and machine learning with applications in smart grids and bioinformatics. Angshul has co-authored over 200 articles in journals and top tier conferences. He has written two books and co-edited two more and holds 7 US patents. He is an associate editor for IEEE Open Journal for Signal Processing and Elsevier Neurocomputing. In the past, he has been an associate editor for IEEE Transactions on Circuits and Systems for Video Technology.
Angshul is currently the director of student services at IEEE Signal Processing Society. Prior to that he was the chair for the education committee in the IEEE SPS membership board (2019). Angshul has also served as the chair for the chapter’s committee in the IEEE SPS membership board (2016-18). He had been the founding chair of IEEE SPS Delhi Chapter (2015-18). Angshul has been the organizing chair of two IEEE SPS Winter Schools in 2014 and 2017. He has served as the finance chair of IEEE ISBA 2017, the flagship conference of IEEE Biometrics Council.
