Authors: Frederik Hogenboom, Marnix Moerland, Michel Capelle
Tags: 2013, conceptual modeling
Traditionally, content-based news recommendation is performed by means of the cosine similarity and the TF-IDF weighting scheme for terms occurring in news messages and user profiles. Semantics-driven variants like SF-IDF additionally take into account term meaning by exploiting synsets from semantic lexicons. However, semantics-based weighting techniques are not able to handle – often crucial – named entities, which are often not present in semantic lexicons. Hence, we extend SF-IDF by also employing named entity similarities using Bing page counts. Our proposed method, Bing-SF-IDF, outperforms TF-IDF and its semantics-driven variants in terms of F 1-scores and kappa statistics.Read the full paper here: https://link.springer.com/chapter/10.1007/978-3-319-14139-8_18
