Developing and Testing Top-N Recommendation Algorithms for 0-1 Data using recommenderlab
Michael Hahsler · 2011
The problem of creating recommendations given a large data base from directly elicited ratings (e.g., ratings of 1 through 5 stars) is a popular research area which was lately boosted by the Netflix Prize competition. While computing recommendations using these type of data has direct application for example for large on-line retailers, there are many potential applications for recommender systems where such data is not available. However, in many cases there might be 0-1 rating data available or can be derived from other data sources (e.g., purchase records, Web click data) which can be utilized. Although this type of data differs significantly from directly elicited ratings, only very limited research is available for 0-1 data. This paper describes recommenderlab which provides the infrastructure to test and develop recommender algorithms. Currently the focus is on recommender systems for 0-1 data, however in the future it can be extended to also the more thoroughly researched case of directly elicited, real-valued rating data. 1