Recommender Systems: An introduction
Chinese version: "Recommendation System"
Authors: Diermar Jannach, Markus Zanker, Alexander Felfernig, Gerhard Friedrich
Main content: This book comprehensively introduces the relevant knowledge points involved in recommender systems, presents many classic algorithms, and discusses how to measure the effectiveness of recommender systems. The content of the book is divided into two parts: basic concepts and recent progress: the former involves collaborative recommendation, content-based recommendation, knowledge-based recommendation, hybrid recommendation methods, explanation of recommendation systems, evaluation of recommendation systems, and case analysis; the latter includes recommendations for recommendation systems. attacks, online consumption decisions, recommender systems and the next-generation Internet, and recommendation in pervasive environments.
Reasons for recommendation: This book is detailed in content, covers a wide range of different types of recommendation systems, and analyzes these recommendation systems in detail one by one, supplemented by the introduction of practical application cases, suitable for those who want to understand the basis of recommendation systems and related research. Readers as introductory books for recommender systems. The book contains a large number of charts and examples to help readers understand and grasp the relevant knowledge.
Recommender Systems: The Textbook
Main content: This book introduces all aspects of recommender systems in detail, and the content is roughly divided into three parts: 1) The "Algorithms and Evaluation" part discusses the basic algorithms in recommender systems, including collaborative filtering methods, content-based methods , knowledge-based methods, ensemble methods, and evaluation methods for recommender systems; 2) The "Domain and Context-Specific Recommender Systems" section introduces the use of different contextual scene data such as spatiotemporal data, social data, label data, and credit data. How to recommend; 3) The "Advanced Topics and Applications" section introduces the robustness of recommender systems, such as shilling systems, attack models, and corresponding defense models.
Recommended reason: This is a very good textbook . It not only explains the basics of recommender systems in concise language, but also introduces the core algorithms and mathematical arguments in depth, and provides readers with a lot of inquiries when using third-party tools or frameworks. material. This book provides a comprehensive introduction to the basics, specific applications, and related literature of recommender systems. It is not only suitable for researchers as an introductory book for recommender systems, but also as a tool reference book for industrial practitioners.
Recommender System Handbook
Main content: This book introduces recommender system technology from two perspectives: 1) basic recommendation algorithms; 2) recommendation system evaluation and application. In terms of basic recommendation technology, the book makes an in-depth summary and analysis of various classic algorithms in the early development of recommendation systems, including content-based recommendation, nearest-neighbor-based collaborative filtering, and matrix factorization. In terms of recommendation system evaluation and application, this book discusses the commonly used methods and criteria for recommendation algorithm evaluation, and introduces the challenges that may be encountered in the implementation of recommendation systems.
Recommended reason: This is a classic book in the field of recommender systems. The book is long, and each chapter has invited well-known scholars to participate in the writing. The classical methods and problems introduced in it are very inspiring for the research and practice of recommendation systems in the era of deep learning. Published in 2011, the 3rd edition was published this year, and is ideal for researchers and engineers as a technical reference manual in actual research and work.
"Recommendation System Practice"
Main content: This is a partial application book compiled by Dr. Xiang Liang from his research and experience on recommendation systems during his doctoral period. It was published in 2012. Combined with specific application scenarios, this book briefly introduces the basic components of recommender systems, and how to flexibly utilize different content data, such as user tags, social networks, contextual information, etc., to help improve recommendation models. The algorithms covered in the book are mainly collaborative filtering, content filtering and graph algorithms. Each chapter has simple algorithm code examples and data result analysis.
Reasons for recommendation: The biggest feature of this book is that it systematically introduces many aspects of recommendation systems from the perspective of practical applications , including typical application scenarios, basic algorithm models, important auxiliary information, evaluation indicators, and recommendation engines. 's architecture. , beginners can understand the basic knowledge of recommendation system in the shortest time through this book, and at the same time have an overview of the basic modules required to build a recommendation system.
"Recommender Systems: Frontiers and Practices"
Main content: This book analyzes the recommendation system for readers from the perspective and experience of front-line R&D personnel, from the perspective of principle and practice. The book first introduces various classic recommendation algorithms and cutting-edge deep learning recommendation algorithms in principle, covering deep collaborative filtering, feature interaction, recommendation based on graph neural network, sequence and session recommendation, recommendation combined with knowledge graph and reinforcement learning based Recommendation and other important technologies, and then discuss the cutting-edge topics of recommendation algorithm research in dialogue, causality, common sense, etc. The book also analyzes the key challenges of recommender systems in data fusion, system scaling, performance evaluation, etc., and discusses how to design responsible recommender systems. Finally, the book introduces the practical experience of recommender systems in conjunction with Microsoft's open source project, Microsoft Recommenders.
Reasons for recommendation: In recent years, the combination of recommendation system and deep learning has been generally recognized by industry and academia, but related books have not generally covered these cutting-edge technologies and practical experience. The authors of this book, Li Dongsheng, Lian Jianxun, Zhang Le, Ren Kan, Lu Tun, Wu Tao, and Xie Xing, have been active in the research and development of recommendation systems for a long time, and have published hundreds of influential articles in authoritative conferences and journals in the field of recommendation systems. Academic papers, and presided over the development of several recommender system projects including Microsoft Recommenders. Based on this book, readers can deeply learn the most cutting-edge recommendation algorithm design principles and practices, and can quickly build an accurate and efficient recommendation system from scratch based on the source code in the book.