260 likes | 377 Vues
This project explores a hybrid recommender system for BoardGameGeek.com, focusing on link analysis and genetic tuning methods in a bipartite network. Utilizing data from the top 5,000 rated board games and active users, the approach includes extracting relevant information using the BGG XML API. We apply genetic algorithms to optimize user and game weights, enhancing the recommendation accuracy. Key methodologies include link analysis techniques, matrix normalization, and a fitness-based genetic tuning process to adapt weights for optimal performance.
E N D
A Hybrid Recommender System Using Link Analysis and Genetic Tuning in the Bipartite Network of BoardGameGeek.com Brett Boge CS 765 University of Nevada, Reno
Data (Scope) • Starting with the top 5,000 games • List of users == those which have rated at least one of the top 5,000 games • Users with no ratings cannot be connected to anycomponent of the graph, and can only be evaluatedin the most general sense
Data (Retrieval) • Data will be obtained through the BGG XML API2 • Game|Small World, id 40692 • http://boardgamegeek.com/xmlapi2/thing?id=40692&ratingcomments=1 • User|Licinian • http://boardgamegeek.com/xmlapi2/user?name=Licinian • http://boardgamegeek.com/xmlapi2/ • collection?name=Licinian • &own/played/trade/want/wishlist/etc
General Approach Approaches R. Burke, "Hybrid recommender systems: Survey and experiments,"
General Approach Approaches
Link Analysis Step Approaches From Z. Huang, et al., "A Link analysis approach to recommendation under sparse data," 2004.
ConsumerRepresentativeness • Matrix Link Analysis Step Approaches • ProductRepresentativeness • Matrix From Z. Huang, et al., "A Link analysis approach to recommendation under sparse data," 2004.
ConsumerRepresentativeness • Matrix Link Analysis Step Approaches • ProductRepresentativeness • Matrix From Z. Huang, et al., "A Link analysis approach to recommendation under sparse data," 2004.
ConsumerRepresentativeness • Matrix Link Analysis Step Approaches • ProductRepresentativeness • Matrix From Z. Huang, et al., "A Link analysis approach to recommendation under sparse data," 2004.
Content-based Cascade Approaches • ProductRepresentativeness • Matrix PRi
Content-based Cascade Approaches
Content-based Cascade Approaches These will need to be normalized on the same scale (0.00 - 1.00)
Content-based Cascade Approaches
Content-based Cascade Approaches • Create PRfinal by refining PR: • W is a vector of weights which determine how much a givenproperty should effect the original score
Genetic Tuning Approaches • W needs to be defined optimally for this given domain • A genetic algorithm will be used to tune W • Chromosome = sequential binary representation of W • Fitness based on Rank Score (from Huang et al.) • 8 bits per weight, ranging from 0 - .25 to start • Rates of crossover/mutation TBD