Paper: Are Semantically Coherent Topic Models Useful for Ad Hoc Information Retrieval?

ACL ID P13-2027
Title Are Semantically Coherent Topic Models Useful for Ad Hoc Information Retrieval?
Venue Annual Meeting of the Association of Computational Linguistics
Session Short Paper
Year 2013
Authors

The current topic modeling approaches for Information Retrieval do not allow to ex- plicitly model query-oriented latent top- ics. More, the semantic coherence of the topics has never been considered in this field. We propose a model-based feedback approach that learns Latent Dirichlet Al- location topic models on the top-ranked pseudo-relevant feedback, and we mea- sure the semantic coherence of those top- ics. We perform a first experimental eval- uation using two major TREC test collec- tions. Results show that retrieval perfor- mances tend to be better when using topics with higher semantic coherence.