Offering Collaborative-like Recommendationis When Data is Sparse: The Case of Attraction-weighted Information Filtering
We propose a low-dimensional weigthing scheme to map information filtering recommendations into more relevant, collaborative filtering-like recommendations where items are weigthed by attraction indexes representing existing customers’ preferences. A first study conducted with consumers within an online bookseller context indicates that recommendations made by our system favorably compare to data-hungry collaborative filtering systems, while requiring much less data.
DE BRUYN, A., GILES, C.L. et PENNOCK, D.M. (2004). Offering Collaborative-like Recommendationis When Data is Sparse: The Case of Attraction-weighted Information Filtering. Dans: Lectures Notes in Computer Science n° 3137. Proceedings of the Third International Conference on Adaptive Hypermedia and Adaptive Web-based Systems. Springer, pp. 393-396.