WISECON The Intelligent Support for Ecommerce Petr Berka
WISECON: The Intelligent Support for E-commerce Petr Berka, Tomáš Kočka, Tomáš Kroupa Laboratory for Intelligent Systems University of Economics, Prague Petr Berka, Tomáš Kroupa, 23. 4. 2003
Intelligent support for Internet shopping n n n help the user to decide which products to buy, find specifications and reviews of the products, make recommendations, find the best price for the desired product (comparison shopping), monitor new products on the product list, watch for special offers or discounts. Petr Berka, Tomáš Kroupa, 23. 4. 2003
Recommender systems “Recommender systems use product knowledge – either hand-coded knowledge provided by experts or “mined” knowledge learned from the behavior of customers – to guide customers through the often-overwhelming task of locating products they will like. ” Schafer J. B. , Konstan J. A. , Riedl, J. : E-Commerce Recommendation Applications. Data Mining and Knowledge Discovery 5 (2001) 115 -153. Petr Berka, Tomáš Kroupa, 23. 4. 2003
E-commerce Recommender application models (1/2) n Broad recommendation list n Helping new and infrequent visitors (no personal info needed) n Customer comments and ratings n Building credibility through community (one-to-one marketing) n Notification services n Inviting customers back Petr Berka, Tomáš Kroupa, 23. 4. 2003
E-commerce Recommender application models (2/2) n Product associated recommendations n n Cross-selling Deep personalization n Building long-term relationships Petr Berka, Tomáš Kroupa, 23. 4. 2003
Example – amazon. com “customer who bought…” n Product associated model (cross-selling) n n n Recommendation method: item-to-item correlation Customer input: implicit navigation Community input: purchase history Output: suggestion Ephemeral personalization Passive delivery of unordered list Petr Berka, Tomáš Kroupa, 23. 4. 2003
Intelligent Shopping Assistant WISECON - overview (1/2) WISECON - Support of access to on-line catalogue of IBM PC’s n Browsing/search n n Clustering products Recommending n n Expert’s knowledge Community input Petr Berka, Tomáš Kroupa, 23. 4. 2003
Intelligent Shopping Assistant WISECON - overview (2/2) n Broad recommendation model n n n Recommendation method: attribute-based Customer input: implicit navigation + keyword/item attributes Community input: purchase history Output: suggestion Personalization: none to ephemeral Passive delivery of ordered list Petr Berka, Tomáš Kroupa, 23. 4. 2003
WISECON Inference Cycle n n n Improve browsing of the on-line catalogue Recommend products Control the communication with the user Petr Berka, Tomáš Kroupa, 23. 4. 2003
Clustering of Products To make both browsing and recommending more comprehensible Petr Berka, Tomáš Kroupa, 23. 4. 2003
Requirements on recommending module n n use expert knowledge easy to update to new products reflect technological development accept vague requests Petr Berka, Tomáš Kroupa, 23. 4. 2003
Possible methods n Bayesian network n n (EUNITE’ 01, ISMIS’ 02) Possibilistic network n (SCI’ 02, IEEE IS’ 02) (expert system, CBR, . . . ) Petr Berka, Tomáš Kroupa, 23. 4. 2003
Possibility vs. Probability Possibility Distribution Marginals Independence Conditional independence Petr Berka, Tomáš Kroupa, 23. 4. 2003 Probability
WISECON Network Petr Berka, Tomáš Kroupa, 23. 4. 2003
Interaction with the User Considering various types of the user Expert n Middle experienced n Inexperienced (this information is given by the user) n Asking only such questions that have the main impact on discrimination between computers selection of questions based on mutual information (in probability) or interactivity measure (in possibility) Petr Berka, Tomáš Kroupa, 23. 4. 2003
- Slides: 15