This paper introduces the paradigm of relatedness, the generalization of the paradigm of user interest. Relatedness is typically interpreted in a graph based information representation environment, where the content-based and collaborative information is treated at the same abstraction level. To demonstrate the effectiveness of the paradigm, various graph based recommendation methods are evaluated on standard datasets, as MovieLens and MovieTweetings. In our experiment, we focus on the information sparse environment and measure coverage, precision, recall and nDCG on top-N recommendation lists. The primary conclusion of our work is that the paradigm of relatedness is a promising direction as the evaluation results show a significant increase in the recommendation quality of the method implementing the paradigm.
CITATION STYLE
Grad-Gyenge, L., & Filzmoser, P. (2017). The paradigm of relatedness. In Lecture Notes in Business Information Processing (Vol. 263, pp. 57–68). Springer Verlag. https://doi.org/10.1007/978-3-319-52464-1_6
Mendeley helps you to discover research relevant for your work.