Towards a scalable kNN CF algorithm: exploring effective applications of clustering

摘要: Collaborative Filtering (CF)-based recommender systems bring mutual benefits to both users and the operators of sites with too much information. Users benefit as they are able find items interest from an unmanageable number available items. On other hand, e-commerce that employ can increase sales revenue in at least two ways: a) by drawing customers' attention likely buy, b) cross-selling However, sheer customers typical demand specially designed CF algorithms gracefully cope vast size data. Many proposed thus far, where principal concern is recommendation quality, may be expensive operate a large-scale system. We propose CLUSTKNN, simple intuitive algorithm well suited for large data sets. The method first compresses tremendously building straightforward but efficient clustering model. Recommendations then generated quickly using NEAREST NEIGHBOR-based approach. demonstrate feasibility CLUSTKNN analytically empirically. also show, comparing popular that, apart being highly scalable intuitive, provides very good accuracy well.