PROCESSES FOR CALCULATING ITEM DISTANCES AND PERFORMING ITEM CLUSTERING

摘要

Computer-implemented processes are disclosed for clustering items and improving the utility of item recommendations. One process (Fig. 2) involves applying a clustering algorithm (block 36) to a user's collection of items. Information about the resulting clusters is then used to select items to use as recommendation sources (block 38) Another process involves displaying the clusters of items to the user via a collection management interface (Fig. 3) that enables the user to attach cluster-level metadata, such as by rating or tagging entire clusters of items. The resulting metadata may be used to improve the recommendations generated by a recommendation engine (62). Another process (Fig. 4) involves using a clustering engine (60) to form clusters of items in which a user has indicated a lack of interest, and using these clusters to filter (64) the output of a recommendation engine (62). Yet another process (Fig. 5) involves applying a clustering algorithm to the output of a recommendation engine (62) to arrange the recommended items into cluster-based categories for presentation to the user (Fig. 7). Also disclosed are methods for calculating distances between items based on item locations in a hierarchical browse structure (Fig. 8), and for using the distance calculations to form clusters.