: Andreas W. Neumann
: Recommender Systems for Information Providers Designing Customer Centric Paths to Information
: Physica-Verlag
: 9783790821345
: 1
: CHF 85.30
:
: Allgemeines, Lexika
: English
: 158
: Wasserzeichen
: PC/MAC/eReader/Tablet
: PDF

Information providers are a very promising application area of recommender systems due to the general problem of assessing the quality of information products prior to the purchase. Recommender systems automatically generate product recommendations: customers profit from a faster finding of relevant products, stores profit from rising sales. All aspects of recommender systems are covered: the economic background, mechanism design, a survey of systems in the Internet, statistical methods and algorithms, service oriented architectures, user interfaces, as well as experiences and data from real-world applications. Specific solutions for areas with strong privacy concerns, scalability issues for large collections of products, as well as algorithms to lessen the cold-start problem for a faster return on investment of recommender projects are addressed. This book describes all steps it takes to design, implement, and successfully operate a recommender system for a specific information platform.

Preface5
Contents7
1 Introduction11
1.1 Recommender Systems11
1.2 Scienti.c and Technical Information14
1.3 Motivation and Focus16
1.4 Chapter Guide18
2 The Market of Scienti.c and Technical Information19
2.1 Information Providers in the Digital Age19
2.2 The 2-3-6-Value-Chain for STI Markets24
2.3 The Strategic Positions of the Market Players26
2.4 Recommender Systems in the Business Pro.le of Information Providers31
3 Classi.cation and Mechanism Design of Recommender Systems32
3.1 Classi.cations of Recommender Systems32
3.2 Mechanism Design35
4 A Survey of Recommender Systems at Major STI Providers40
4.1 Scienti.c Libraries40
4.2 Scienti.c Projects44
4.3 E-Commerce: Amazon.com45
4.4 Social Tagging48
5 Case Study: Explicit Recommender Services for Scienti.c Libraries52
5.1 General Setup52
5.2 Rating Service56
5.3 Review Service57
5.4 Usage Statistics59
5.5 Discussion60
6 General Concepts of Behavior-Based Recommender Services65
6.1 Revealed Preference Theory and Choice Sets65
6.2 Self-Selection67
6.3 Prices, Transaction Costs, Market Baskets, Lending Data, and Browser Sessions67
6.4 Knowledge Discovery and Data Mining69
6.5 Observed Users vs. Target Group of Recommendations69
6.6 Factors for the Di.usion of Recommendations70
7 Algorithms for Behavior-Based Recommender Systems72
7.1 Purchase Noise Filtering by Means of the Logarithmic Series Distribution72
7.2 POSICI and POMICI: Recommendations from Small Samples83
7.3 Related Methods94
8 Case Study: Behavior-Based Recommender Services for Scienti.c Libraries97
8.1 Service Description97
8.2 Implementation103
8.3 Evaluation112
8.4 BibTip: Commercial Implicit Recommendation Services for Libraries121
8.5 Extensions, Improvements, and Further Applications122
9 Visualizing and Exploring Information Spaces126
9.1 A Survey of Visual Interfaces to Information Providers127
9.2 RecoDiver: A Graph-Based User Interface to Recommendations129
9.3 Evaluation135
9.4 Discussion and Outlook137
10 Discussion139
List of Figures143
List of Tables146
References147