You are previewing Recommender Systems.
O'Reilly logo
Recommender Systems

Book Description

In this age of information overload, people use a variety of strategies to make choices about what to buy, how to spend their leisure time, and even whom to date. Recommender systems automate some of these strategies with the goal of providing affordable, personal, and high-quality recommendations. This book offers an overview of approaches to developing state-of-the-art recommender systems. The authors present current algorithmic approaches for generating personalized buying proposals, such as collaborative and content-based filtering, as well as more interactive and knowledge-based approaches. They also discuss how to measure the effectiveness of recommender systems and illustrate the methods with practical case studies. The final chapters cover emerging topics such as recommender systems in the social web and consumer buying behavior theory. Suitable for computer science researchers and students interested in getting an overview of the field, this book will also be useful for professionals looking for the right technology to build real-world recommender systems.

Table of Contents

  1. Cover
  2. Half-title
  3. Title
  4. Copyright
  5. Contents
  6. Foreword
  7. Preface
  8. 1 Introduction
    1. 1.1 Part I: Introduction to basic concepts
      1. 1.1.1 Collaborative recommendation
      2. 1.1.2 Content-based recommendation
      3. 1.1.3 Knowledge-based recommendation
      4. 1.1.4 Hybrid approaches
      5. 1.1.5 Explanations in recommender systems
      6. 1.1.6 Evaluating recommender systems
      7. 1.1.7 Case study
    2. 1.2 Part II: Recent developments
  9. PART I: Introduction to basic concepts
    1. 2 Collaborative recommendation
      1. 2.1 User-based nearest neighbor recommendation
        1. 2.1.1 First example
        2. 2.1.2 Better similarity and weighting metrics
        3. 2.1.3 Neighborhood selection
      2. 2.2 Item-based nearest neighbor recommendation
        1. 2.2.1 The cosine similarity measure
        2. 2.2.2 Preprocessing data for item-based filtering
      3. 2.3 About ratings
        1. 2.3.1 Implicit and explicit ratings
        2. 2.3.2 Data sparsity and the cold-start problem
      4. 2.4 Further model-based and preprocessing-based approaches
        1. 2.4.1 Matrix factorization/latent factor models
        2. 2.4.2 Association rule mining
        3. 2.4.3 Probabilistic recommendation approaches
      5. 2.5 Recent practical approaches and systems
        1. 2.5.1 Slope One predictors
        2. 2.5.2 The Google News personalization engine
      6. 2.6 Discussion and summary
      7. 2.7 Bibliographical notes
    2. 3 Content-based recommendation
      1. 3.1 Content representation and content similarity
        1. 3.1.1 The vector space model and TF-IDF
        2. 3.1.2 Improving the vector space model/limitations
      2. 3.2 Similarity-based retrieval
        1. 3.2.1 Nearest neighbors
        2. 3.2.2 Relevance feedback – Rocchio’s method
      3. 3.3 Other text classification methods
        1. 3.3.1 Probabilistic methods
        2. 3.3.2 Other linear classifiers and machine learning
        3. 3.3.3 Explicit decision models
        4. 3.3.4 On feature selection
      4. 3.4 Discussion
        1. 3.4.1 Comparative evaluation
        2. 3.4.2 Limitations
      5. 3.5 Summary
      6. 3.6 Bibliographical notes
    3. 4 Knowledge-based recommendation
      1. 4.1 Introduction
      2. 4.2 Knowledge representation and reasoning
        1. 4.2.1 Constraints
        2. 4.2.2 Cases and similarities
      3. 4.3 Interacting with constraint-based recommenders
        1. 4.3.1 Defaults
        2. 4.3.2 Dealing with unsatisfiable requirements and empty result sets
        3. 4.3.3 Proposing repairs for unsatisfiable requirements
        4. 4.3.4 Ranking the items/utility-based recommendation
      4. 4.4 Interacting with case-based recommenders
        1. 4.4.1 Critiquing
        2. 4.4.2 Compound critiquing
        3. 4.4.3 Dynamic critiquing
        4. 4.4.4 Advanced item recommendation
        5. 4.4.5 Critique diversity
      5. 4.5 Example applications
        1. 4.5.1 The VITA constraint-based recommender
        2. 4.5.2 The Entree case-based recommender
      6. 4.6 Bibliographical notes
    4. 5 Hybrid recommendation approaches
      1. 5.1 Opportunities for hybridization
        1. 5.1.1 Recommendation paradigms
        2. 5.1.2 Hybridization designs
      2. 5.2 Monolithic hybridization design
        1. 5.2.1 Feature combination hybrids
        2. 5.2.2 Feature augmentation hybrids
      3. 5.3 Parallelized hybridization design
        1. 5.3.1 Mixed hybrids
        2. 5.3.2 Weighted hybrids
        3. 5.3.3 Switching hybrids
      4. 5.4 Pipelined hybridization design
        1. 5.4.1 Cascade hybrids
        2. 5.4.2 Meta-level hybrids
      5. 5.5 Discussion and summary
      6. 5.6 Bibliographical notes
    5. 6 Explanations in recommender systems
      1. 6.1 Introduction
      2. 6.2 Explanations in constraint-based recommenders
        1. 6.2.1 Example
        2. 6.2.2 Generating explanations by abduction
        3. 6.2.3 Analysis and outline of well-founded explanations
        4. 6.2.4 Well-founded explanations
      3. 6.3 Explanations in case-based recommenders
      4. 6.4 Explanations in collaborative filtering recommenders
      5. 6.5 Summary
    6. 7 Evaluating recommender systems
      1. 7.1 Introduction
      2. 7.2 General properties of evaluation research
        1. 7.2.1 General remarks
        2. 7.2.2 Subjects of evaluation design
        3. 7.2.3 Research methods
        4. 7.2.4 Evaluation settings
      3. 7.3 Popular evaluation designs
      4. 7.4 Evaluation on historical datasets
        1. 7.4.1 Methodology
        2. 7.4.2 Metrics
        3. 7.4.3 Analysis of results
      5. 7.5 Alternate evaluation designs
        1. 7.5.1 Experimental research designs
        2. 7.5.2 Quasi-experimental research designs
        3. 7.5.3 Nonexperimental research designs
      6. 7.6 Summary
      7. 7.7 Bibliographical notes
    7. 8 Case study: Personalized game recommendations on the mobile Internet
      1. 8.1 Application and personalization overview
      2. 8.2 Algorithms and ratings
      3. 8.3 Evaluation
        1. 8.3.1 Measurement 1: “My Recommendations”
        2. 8.3.2 Measurement 2: Post-sales recommendations
        3. 8.3.3 Measurement 3: Start page recommendations
        4. 8.3.4 Measurement 4: Overall effect on demo downloads
        5. 8.3.5 Measurement 5: Overall effects
      4. 8.4 Summary and conclusions
  10. PART II: Recent developments
    1. 9 Attacks on collaborative recommender systems
      1. 9.1 A first example
      2. 9.2 Attack dimensions
      3. 9.3 Attack types
        1. 9.3.1 The random attack
        2. 9.3.2 The average attack
        3. 9.3.3 The bandwagon attack
        4. 9.3.4 The segment attack
        5. 9.3.5 Special nuke attacks
        6. 9.3.6 Clickstream attacks and implicit feedback
      4. 9.4 Evaluation of effectiveness and countermeasures
        1. 9.4.1 Push attacks
        2. 9.4.2 Nuke attacks
      5. 9.5 Countermeasures
      6. 9.6 Privacy aspects -- distributed collaborative filtering
        1. 9.6.1 Centralized methods: Data perturbation
        2. 9.6.2 Distributed collaborative filtering
      7. 9.7 Discussion
    2. 10 Online consumer decision making
      1. 10.1 Introduction
      2. 10.2 Context effects
      3. 10.3 Primacy/recency effects
      4. 10.4 Further effects
      5. 10.5 Personality and social psychology
      6. 10.6 Bibliographical notes
    3. 11 Recommender systems and the next-generation web
      1. 11.1 Trust-aware recommender systems
        1. 11.1.1 Exploiting explicit trust networks
        2. 11.1.2 Trust metrics and effectiveness
        3. 11.1.3 Related approaches and recent developments
      2. 11.2 Folksonomies and more
        1. 11.2.1 Using folksonomies for recommendations
          1. 11.2.1.1 Folksonomies and content-based methods
          2. 11.2.1.2 Folksonomies and collaborative filtering
        2. 11.2.2 Recommending tags
        3. 11.2.3 Recommending content in participatory media
      3. 11.3 Ontological filtering
        1. 11.3.1 Augmentation of filtering by taxonomies
        2. 11.3.2 Augmentation of filtering by attributes
      4. 11.4 Extracting semantics from the web
      5. 11.5 Summary
    4. 12 Recommendations in ubiquitous environments
      1. 12.1 Introduction
      2. 12.2 Context-aware recommendation
      3. 12.3 Application domains
      4. 12.4 Summary
    5. 13 Summary and outlook
      1. 13.1 Summary
      2. 13.2 Outlook
  11. Bibliography
  12. Index