Cross-lingual keyword recommendation using latent topics

Abstract

Multi-lingual text processing is important for content-based and hybrid recommender systems. It helps recommender systems extract content information from broader sources. It also enables systems to recommend items in a user's native language. We propose a cross-lingual keyword recommendation method, which is built on an extended latent Dirichlet allocation model, for extracting latent features from parallel corpora. With this model, the proposed method can recommend keywords from text written in different languages. We evaluate the proposed method using a cross-lingual bibliographic database that contains both English and Japanese abstracts and keywords and show that the proposed method can recommend keywords from abstracts in a cross-lingual environment with almost the same accuracy as in a monolingual environment. © 2010 ACM.