2.3.4 Bayesian Classifiers 2.3.5 Artificial Neural Networks 2.3.6 Support Vector Machines 2.3.7 Ensembles of Classifiers 2.3.8 Evaluating Classifiers 2.4 Cluster Analysis 2. 4. 1 k-Means 2.4.2 Alternatives to k-means 2.5 Association Rule Mining 2.6 Conclusions References 3 Content-based Recommender Systems: State of the Art and Trends. 73 Pasquale Lops, Marco de Gemmis and Giovanni Semeraro 3.1 Introduction 3.2 Basics of Content-based Recommender Systems 75 3.2. 1 A High Level Architecture of Content-based Systems.. 75 3.2.2 Advantages and Drawbacks of Content-based Filtering.. 78 3.3 State of the Art of Content-based Recommender Systems 3.3.1 Item Representation 3.3.2 Methods for Learning user profiles 3.4 Trends and Future research 3.4.1 The Role of User Generated Content in the Recommendation process 3.4.2 Beyond Over-specializion: Serendipity 3.5 Conclusions References 4 A Comprehensive Survey of Neighborhood-based Recommendation methods Christian Desrosiers and George Karypis 4.1.1 Formal Definition of the problem 4. 1.2 Overview of Recommendation Approaches 110 1.3 Advantages of Neighborhood Approaches 4. 1. 4 Objectives and Outline 4.2 Neighborhood-based Recommendation 114 4.2.1 User-based Rating prediction 4.2.2 User-based Classification 116 egression Vs Classification 4. 2. 4 Item-based Recommendation 117 4.2.5 User-based VS Item-based Recommendatio 118 ponents of Neighborhood Methods 4.3.1 Rating normalization 121 4.3.2 Similarity Weight Computation 4.3.3 Neighborhood Selection
x Contents 2.3.4 Bayesian Classifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.3.5 Artificial Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . . 54 2.3.6 Support Vector Machines . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 2.3.7 Ensembles of Classifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 2.3.8 Evaluating Classifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 2.4 Cluster Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 2.4.1 k-Means . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.4.2 Alternatives to k-means . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 2.5 Association Rule Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 3 Content-based Recommender Systems: State of the Art and Trends . 73 Pasquale Lops, Marco de Gemmis and Giovanni Semeraro 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 3.2 Basics of Content-based Recommender Systems . . . . . . . . . . . . . . . 75 3.2.1 A High Level Architecture of Content-based Systems . . . 75 3.2.2 Advantages and Drawbacks of Content-based Filtering . . 78 3.3 State of the Art of Content-based Recommender Systems . . . . . . . . 79 3.3.1 Item Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 3.3.2 Methods for Learning User Profiles . . . . . . . . . . . . . . . . . . 90 3.4 Trends and Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 3.4.1 The Role of User Generated Content in the Recommendation Process . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 3.4.2 Beyond Over-specializion: Serendipity . . . . . . . . . . . . . . . . 96 3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 4 A Comprehensive Survey of Neighborhood-based Recommendation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Christian Desrosiers and George Karypis 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 4.1.1 Formal Definition of the Problem . . . . . . . . . . . . . . . . . . . . 108 4.1.2 Overview of Recommendation Approaches . . . . . . . . . . . . 110 4.1.3 Advantages of Neighborhood Approaches . . . . . . . . . . . . . 112 4.1.4 Objectives and Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 4.2 Neighborhood-based Recommendation . . . . . . . . . . . . . . . . . . . . . . . 114 4.2.1 User-based Rating Prediction . . . . . . . . . . . . . . . . . . . . . . . . 115 4.2.2 User-based Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 4.2.3 Regression VS Classification . . . . . . . . . . . . . . . . . . . . . . . . 117 4.2.4 Item-based Recommendation . . . . . . . . . . . . . . . . . . . . . . . . 117 4.2.5 User-based VS Item-based Recommendation . . . . . . . . . . 118 4.3 Components of Neighborhood Methods. . . . . . . . . . . . . . . . . . . . . . . 120 4.3.1 Rating Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 4.3.2 Similarity Weight Computation . . . . . . . . . . . . . . . . . . . . . . 124 4.3.3 Neighborhood Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 . . . . . . . .
Contents 4.4 Advanced Techniques 4.4. 1 Dimensionality Reduction Methods 132 4.4.2 Graph-based Methods 135 4.5 Conclusion References 140 5 Advances in Collaborative Filtering 145 Yehuda Koren and robert bell 5.1 Introduction 145 5.2 Preliminaries 147 5.2. 1 Baseline predictor 5.2.2 The Netflix data 149 5.2.3 Implicit feedback 5.3 Matrix factorization models 5.3.1sVD 6. 2 SVD+ 5.3.3 Time-aware factor model 5.3.4 Comparison 5.3.5 Summary 5.4 Neighborhood models 5.4.1 Simila 5.4.2 Similarity-based interpolation 163 5.4.3 Jointly derived interpolation weights 5.5 Enriching neighborhood models 5.5. 1 A global neighborhood model 5.5.2 A factorized neighborhood model 173 5.5.3 Temporal dynamics at neighborhood models 182 5.6 Between neighborhood and factorization References 6 Developing Constraint-based Recommenders 187 Alexander Felfernig, Gerhard Friedrich, Dietmar Jann Markus zanker 6.1 Introduction 6.2 Development of Recommender Knowledge Bases 6.3 User Guidance in Recommendation Processes 6. 4 Calculating recommendations 203 6.5 Experiences from Projects and Case Studies 6.6 Future Research issues 212 References 212
Contents xi 4.4 Advanced Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 4.4.1 Dimensionality Reduction Methods . . . . . . . . . . . . . . . . . . 132 4.4.2 Graph-based Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 5 Advances in Collaborative Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Yehuda Koren and Robert Bell 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 5.2 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 5.2.1 Baseline predictors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 5.2.2 The Netflix data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 5.2.3 Implicit feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 5.3 Matrix factorization models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 5.3.1 SVD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 5.3.2 SVD++ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 5.3.3 Time-aware factor model . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 5.3.4 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 5.3.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 5.4 Neighborhood models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 5.4.1 Similarity measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 5.4.2 Similarity-based interpolation . . . . . . . . . . . . . . . . . . . . . . . 163 5.4.3 Jointly derived interpolation weights . . . . . . . . . . . . . . . . . 165 5.4.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 5.5 Enriching neighborhood models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 5.5.1 A global neighborhood model . . . . . . . . . . . . . . . . . . . . . . . 169 5.5.2 A factorized neighborhood model . . . . . . . . . . . . . . . . . . . . 173 5.5.3 Temporal dynamics at neighborhood models . . . . . . . . . . . 180 5.5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 5.6 Between neighborhood and factorization . . . . . . . . . . . . . . . . . . . . . . 182 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 6 Developing Constraint-based Recommenders . . . . . . . . . . . . . . . . . . . . 187 Alexander Felfernig, Gerhard Friedrich, Dietmar Jannach and Markus Zanker 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 6.2 Development of Recommender Knowledge Bases . . . . . . . . . . . . . . 191 6.3 User Guidance in Recommendation Processes . . . . . . . . . . . . . . . . . 194 6.4 Calculating Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 6.5 Experiences from Projects and Case Studies . . . . . . . . . . . . . . . . . . . 205 6.6 Future Research Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 6.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
7 Context-Aware Recommender Systems 217 Gediminas Adomavicius and Alexander Tuzhilin 7.1 Introduction and motivation 218 7.2 Context in Recommender Systems 7. 2.1 What is Context? 219 7. 2.2 Modeling Contextual Information in 7. 2.3 Obtaining Contextual Information 7.3 Paradigms for Incorporating Context in Recommender Systems .. 230 .3.1 Contextual Pre-Filtering 233 7.3.2 Contextual Post-Filtering 7.3.3 Contextual Modeling 7.4 Combining Multiple Approaches 243 7.4.1 Case Study of Combining Multiple Pre-Filters Algorithms 44 7.4.2 Case Study of Combining Multiple Pre-Filters Experimental Results 245 7.5 Additional Issues in Context-Aware Recommender Systems 247 7.6 Conclusions References Part II Applications and Evaluation of rss Evaluating Recommendation Systems Guy Shani and Asela Gunawardana 8.1 Introduction 8. 2.2 User Studies 8.2.3 8. 2.4 Drawing reliable conclusions 3 Recommendation System Properties 8.3.1 User Preference 272 8.3.2 Prediction Accuracy 273 281 8.3.4 Confidence 8.3.5 Trust 8.3.6 Novelty 8.3.7 Serendipity 86 8.3.9 Utilit 8.3.10Risk 8.3.11 Robustness 8.3.12 Privac 291
xii Contents 7 Context-Aware Recommender Systems . . . . . . . . . . . . . . . . . . . . . . . . . 217 Gediminas Adomavicius and Alexander Tuzhilin 7.1 Introduction and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 7.2 Context in Recommender Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 7.2.1 What is Context? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 7.2.2 Modeling Contextual Information in Recommender Systems . . . . . . . . . . . . . . . . . . . . . . . . 223 7.2.3 Obtaining Contextual Information . . . . . . . . . . . . . . . . . . . . 228 7.3 Paradigms for Incorporating Context in Recommender Systems . . 230 7.3.1 Contextual Pre-Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 7.3.2 Contextual Post-Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 7.3.3 Contextual Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 7.4 Combining Multiple Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 7.4.1 Case Study of Combining Multiple Pre-Filters: Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 7.4.2 Case Study of Combining Multiple Pre-Filters: Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 7.5 Additional Issues in Context-Aware Recommender Systems . . . . . 247 7.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 Part II Applications and Evaluation of RSs 8 Evaluating Recommendation Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 257 Guy Shani and Asela Gunawardana 8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258 8.2 Experimental Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 8.2.1 Offline Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 8.2.2 User Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 8.2.3 Online Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 8.2.4 Drawing Reliable Conclusions . . . . . . . . . . . . . . . . . . . . . . . 267 8.3 Recommendation System Properties . . . . . . . . . . . . . . . . . . . . . . . . . 271 8.3.1 User Preference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 8.3.2 Prediction Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 8.3.3 Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281 8.3.4 Confidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 8.3.5 Trust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 8.3.6 Novelty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285 8.3.7 Serendipity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 8.3.8 Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 8.3.9 Utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 8.3.10 Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 8.3.11 Robustness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 8.3.12 Privacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 8.3.13 Adaptivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292 . . .
Contents 8.3.14 Scalability 293 8.4 Conclusion References 9 A Recommender System for an IPTV Service Provider: a Real Large-Scale Production Environment Riccardo bambini. Paolo cremonesi and roberto turrin 9.1 Introduction 9.2 IPTV Architecture 9.2.1 IPTV Search Problems 9.3 Recommender System Architecture 9.3.1 Data Collection 9.3.2 Batch and Real-Time Stages 306 9.4 Recommender Algorithms 9.4.1 Overview of Recommender Algorithms 9.4.2 LSA Content-Based Algorithm 311 9.4.3 Item-based Collaborative algorithm 314 9.4.4 Dimensionality-Reduction-Based Collaborative 9.5 Recommender Services 18 9.6 System Evaluation 9.6.1 Off-Line Analysis 321 9.6.2 On-line Analysis 9.7 Conclusions References 0 How to Get the recommender out of the lab? 333 Jerome Picault, Myriam Ribiere, David Bonnefoy and Kevin Mercer 10.1 Introduction 334 10.2 Designing Real-World Recommender Systems 10.3 Understanding the recommender environment 335 0.3.1 Application Model 10.3.2 User Model 10.3.3 Data Model 344 10.4 Understanding the Recommender validation Steps in an Iterative 4g 10.3.4 A Method for Using Environment Models Design Process 0.41 Validation of the algorithms 350 10.4.2 Validation of the recommendations 351 10.5 Use Case: a Semantic News Recommendation System 10.5.1 Context: MESH Project 356 10.5.2 Environmental models in mesh 357 10.5.3 In Practice: Iterative Instantiations of models 10.6 Conclusion References
Contents xiii 8.3.14 Scalability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 8.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294 9 A Recommender System for an IPTV Service Provider: a Real Large-Scale Production Environment. . . . . . . . . . . . . . . . . . . . . . . . . . . 299 Riccardo Bambini, Paolo Cremonesi and Roberto Turrin 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 9.2 IPTV Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 9.2.1 IPTV Search Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302 9.3 Recommender System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 303 9.3.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 9.3.2 Batch and Real-Time Stages . . . . . . . . . . . . . . . . . . . . . . . . 306 9.4 Recommender Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 9.4.1 Overview of Recommender Algorithms . . . . . . . . . . . . . . . 308 9.4.2 LSA Content-Based Algorithm . . . . . . . . . . . . . . . . . . . . . . 311 9.4.3 Item-based Collaborative Algorithm . . . . . . . . . . . . . . . . . . 314 9.4.4 Dimensionality-Reduction-Based Collaborative Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316 9.5 Recommender Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 9.6 System Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 9.6.1 Off-Line Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 9.6.2 On-line Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325 9.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 10 How to Get the Recommender Out of the Lab? . . . . . . . . . . . . . . . . . . 333 Jerome Picault, Myriam Ribi ´ ere, David Bonnefoy and Kevin Mercer ` 10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334 10.2 Designing Real-World Recommender Systems . . . . . . . . . . . . . . . . . 334 10.3 Understanding the Recommender Environment . . . . . . . . . . . . . . . . 335 10.3.1 Application Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335 10.3.2 User Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340 10.3.3 Data Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344 10.3.4 A Method for Using Environment Models . . . . . . . . . . . . . 349 10.4 Understanding the Recommender Validation Steps in an Iterative Design Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350 10.4.1 Validation of the Algorithms . . . . . . . . . . . . . . . . . . . . . . . . 350 10.4.2 Validation of the Recommendations . . . . . . . . . . . . . . . . . . 351 10.5 Use Case: a Semantic News Recommendation System . . . . . . . . . . 355 10.5.1 Context: MESH Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356 10.5.2 Environmental Models in MESH. . . . . . . . . . . . . . . . . . . . . 357 10.5.3 In Practice: Iterative Instantiations of Models . . . . . . . . . . 361 10.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
Contents ll Matching Recommendation Technologies and Domains 367 Robin Burke and Maryam Ramezani 11.1 Introduction 367 11.2 Related Work 11.3 Knowledge Sources 11.3.1 Recommendation types 370 11.4 Domain 11.4.1H eity 372 11.4.2Risk 1143 Churn 373 11.4.4 Interaction Style 374 11. 4.5 Preference stability 11.4.6 Scrutability 375 11.5 Knowledge Sources 11.5.1 Social Knowledge 375 11.5.2 Individual 11.5.3 Content 377 11.6 Mapping Domains to Technologies 11.6.1 Algorithms 11.6.2 Sample Recommendation Domains 381 11.7 Conclusion References 12 Recommender Systems in Technology Enhanced Learning 387 Nikos manouselis Hendrik Drachsler. Riina Vuorikari Hans Hummel nd Rob Koper 12.1 Introduction 12.2 Background 12. 3 Related Work 392 12.4 Survey of TEL Recommender Systems 12.5 Evaluation of TEL Recommender 12.6 Conclusions and further work 408 References Part Ill Interacting with Recommender Systems 13 On the Evolution of Critiquing Recommenders Lorraine McGinty and James Reilly 13.1 Introduction 419 13.2 The Early Days: Critiquing Systems/Recognised Benefits 13.3 Representation Retrieval Challenges for Critiquing Systems .. 422 13.3.1 Approaches to Critique Representation 13.3.2 Retrieval Challenges in Critique-Based Recommenders. 430 13.4 Interfacing Considerations Across Critiquing Platforms 438 13.4.1 Scaling to Alternate Critiquing Platforms 13.4.2 Direct Manipulation Interfaces vs Restricted User Control
xiv Contents 11 Matching Recommendation Technologies and Domains . . . . . . . . . . . 367 Robin Burke and Maryam Ramezani 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367 11.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368 11.3 Knowledge Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368 11.3.1 Recommendation types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370 11.4 Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372 11.4.1 Heterogeneity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372 11.4.2 Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 11.4.3 Churn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373 11.4.4 Interaction Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374 11.4.5 Preference stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374 11.4.6 Scrutability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375 11.5 Knowledge Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375 11.5.1 Social Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375 11.5.2 Individual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376 11.5.3 Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377 11.6 Mapping Domains to Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . 378 11.6.1 Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380 11.6.2 Sample Recommendation Domains. . . . . . . . . . . . . . . . . . . 381 11.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382 12 Recommender Systems in Technology Enhanced Learning . . . . . . . . . 387 Nikos Manouselis, Hendrik Drachsler, Riina Vuorikari, Hans Hummel and Rob Koper 12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388 12.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389 12.3 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392 12.4 Survey of TEL Recommender Systems . . . . . . . . . . . . . . . . . . . . . . . 399 12.5 Evaluation of TEL Recommenders . . . . . . . . . . . . . . . . . . . . . . . . . . . 404 12.6 Conclusions and further work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409 Part III Interacting with Recommender Systems 13 On the Evolution of Critiquing Recommenders . . . . . . . . . . . . . . . . . . 419 Lorraine McGinty and James Reilly 13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419 13.2 The Early Days: Critiquing Systems/Recognised Benefits . . . . . . . . 420 13.3 Representation & Retrieval Challenges for Critiquing Systems . . . 422 13.3.1 Approaches to Critique Representation. . . . . . . . . . . . . . . . 422 13.3.2 Retrieval Challenges in Critique-Based Recommenders . . 430 13.4 Interfacing Considerations Across Critiquing Platforms . . . . . . . . . 438 13.4.1 Scaling to Alternate Critiquing Platforms . . . . . . . . . . . . . . 438 13.4.2 Direct Manipulation Interfaces vs Restricted User Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 440