A topic inference based translation model for question retrieval in community-based question answering services

The ranking scheme of the statistical translation based question retrieval models is mainly depended on the translation probabilities between terms. However, the existing translation based models yield on the noise generated by the translation model and further impact the question retrieval results. In this paper, we proposed a topic inference based translation model for question retrieval. By leveraging the topic information, we theoretically verified that it can reasonably control the translation noise and then improves the question retrieval results. Experimental results show that the proposed model significantly outperforms the state-of-the-art question retrieval models in MAP (Mean Average Precision), MRR (Mean Reciprocal Rank) and p@1 (precision at position one).