Structured Models for Fine-to-Coarse Sentiment Analysis

This Paper is reviewed for Social Media Analysis 10-802 in Fall 2012.

Citation

Ryan Mcdonald , Kerry Hannan , Tyler Neylon , Mike Wells , Jeff Reynar, 2007, In Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics .

Online version

Structured Models for Fine-to-Coarse Sentiment Analysis

Summary

This paper proposes a novel approach to finding sentiments at several granular levels (document, paragraph, sentence, phrase, word ). This paper introduces a single structured model that transforms the multi-level sentiment classification task into a a single problem of learning from sequence of granular components using constrained viterbi. Single model approach performs better than models trained in isolation for a given level of granularity. It considers two important ideas in modelling, firstly, higher level classification can benefit from granular level classification and secondly, granular level classification can benefit from higher level classification.

Results

The results are given for two levels of granularity (sentence and document). Three baseline systems for comparison

1. Document-classifier learns to predict document label only.
2. Sentence-classifier learns to predict isolated sentence label only.
3. Sentence-Structured learns to predict the sentence label by considering the document as a sequence of sentences and uses sequence chain model.

Two alternative to fine-to-coarse systems ( cascaded models ) for comparison

1. Sentence-Structured Model->Document Classifier
2. Document Classifier->Sentence-Structured Model

The proposed model is called Joint-Structured.