Semi-supervised latent variable models for sentence-level sentiment analysis

Täckström, Oscar and McDonald, Ryan (2011) Semi-supervised latent variable models for sentence-level sentiment analysis. In: The 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, 19-24 June 2011, Portland, Oregon, USA. (In Press)

PDF - Accepted Version


We derive two variants of a semi-supervised model for fine-grained sentiment analysis. Both models leverage abundant natural supervision in the form of review ratings, as well as a small amount of manually crafted sentence labels, to learn sentence-level sentiment classifiers. The proposed model is a fusion of a fully supervised structured conditional model and its partially supervised counterpart. This allows for highly efficient estimation and inference algorithms with rich feature definitions. We describe the two variants as well as their component models and verify experimentally that both variants give significantly improved results for sentence-level sentiment analysis compared to all baselines.

Item Type:Conference or Workshop Item (Poster)
ID Code:4157
Deposited By:Oscar Tackström
Deposited On:27 Apr 2011 17:08
Last Modified:27 Apr 2011 17:08

Repository Staff Only: item control page