mining semantic data for solving first-rater and cold-start problems in recommender systems

Download Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender Systems

Post on 24-Feb-2016

35 views

Category:

Documents

0 download

Embed Size (px)

DESCRIPTION

IDEAS 2011 Lisbon 21-23 September. Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender Systems. Data Mining Research Group http://mida.usal.es. María N. Moreno, Saddys Segrera , Vivian F. López, M. Dolores Muñoz and Ángel Luis Sánchez. Department of - PowerPoint PPT Presentation

TRANSCRIPT

hola

1Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis Snchez

Department of Computing and Automatic

Data Mining Research Grouphttp://mida.usal.es

IDEAS 2011Lisbon21-23 SeptemberCEDI 20101ContentsIntroductionRecommender SystemsRecommendation frameworkCase StudyConclusionsMining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 20102Introduction

ClientServer

Catalog

commerce

InformationRecommender systems

Applications: e-commerce, e-learning, tourism, news pagesDrawbacks: low performance, low reliability of recommendations

Recommender systems provide users with intelligent mechanisms to find products to purchaseMining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 20103IntroductionProposalObjective: overcome critical drawbacks in recommender systemsMethodology: Semantic based Web MiningAssociative classification (Web Mining)Machine learning technique that combines concepts from classification and associationDomain-specific ontology (Semantic Web)Enrichment of the data to be mined with semantic annotationsMining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 20104Recommender SystemsClassification of recommendation methodsContent-based: compare text documents to user profilesCollaborative filtering: is based on opinions of other users (ratings)Memory based (User-based): find users with similar preferences (neighbors) by means of statistical techniques Model based (Item-based): use data mining techniques to develop a model of user ratings

Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 20105Recommender SystemsCritical drawbacksSparsity: the number of ratings needed for prediction is greater than the number of the ratings obtained from usersScalability: performance problems presented mainly in memory-based methods where the computation time grows linearly with both the number of customers and the number of products in the siteFirst-rater problem: new products never have been rated, therefore they cannot be recommendedCold-Start problem: new users cannot receive recommendations since they have no evaluations about products

Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 20106Recommendation frameworkAssociative classification (Web Mining)Sparsity: slightly sensitive to sparse dataScalability: model based approachDomain-specific ontology (Semantic Web)First-rater problem: Use of taxonomies to classify productsInduction of abstracts patterns which relate user profiles with categories of productsCold-Start problem: Recommendations based on user profilesMining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 20107Recommendation framework

Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 20108Case Study

User Data

ZipNum. Movies Data

TitleStringIDNum.Genre (19 attributes)BinaryIDNum. Ratings DataRatingNum. (1 - 5)Movie IDNum.User IDNum.IDNum.OccupationStringGenderBinaryAgeNum.

MovieLens DataCEDI 20109Case StudyIDNum.User GenderBinary*User Age < 18 [18, 24] [25, 34] [35, 44] [45, 49] [50, 55] > 55User OccupationStringMovie TitleString*Movie Genre String

MovieLens DataCEDI 201010Case StudyOntology definition

Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 201011Case StudyResultsAssociative classification methods (CBA, CMAR, FOIL and CPAR) were compared to non-associative classification algorithms

Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 201012ConclusionsA framework for recommender systems is proposed in order to overcome some critical drawbacksThe proposal combines web mining methods and domain specific ontologies in order to induce models at two abstraction levels:The low level model relates users, movies and ratings for making the recommendationsHigh level model is used for recommender not rated movies or for making recommendation to new users and overcome the first-rater and the cold-start problemThe off-line model induction avoids scalability problems in recommendation timeAssociative classification methods provides a way to deal with sparsity problemMining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz and ngel Luis SnchezCEDI 201013THANKS FOR YOUR ATTENTION !Mining Semantic Data for Solving First-rater and Cold-start Problems in Recommender SystemsMara N. Moreno*, Saddys Segrera, Vivian F. Lpez, M. Dolores Muoz & ngel Luis Snchez*mmg@usal.es

Department of Computing and Automatic

IDEAS 2011Lisbon21-23 SeptemberCEDI 201014

Recommended

View more >