Fine-grained Opinion Mining with Recurrent Neural Networks and Word Embeddings

Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1433–1443, Lisbon, Portugal, 17-21 September 2015. c⃝2015 Association for Computational Linguistics. Fine-grained Opinion Mining with Recurrent Neural Networks and Word Embeddings Pengfei Liu1, Shafiq Joty2 and Helen Meng1 1Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong, Hong Kong SAR, China 2Qatar Computing Research Institute - HBKU, Doha, Qatar {pfliu, hmmeng}@se.cuhk.edu.hk, sjoty@qf.org.qa Abstract The tasks in fine-grained opinion mining can be regarded as either a token-level se- quence labeling problem or as a semantic compositional task. We propose a gen- eral class of discriminative models based on recurrent neural networks (RNNs) and word embeddings that can be successfully applied to such tasks without any task- specific feature engineering effort. Our experimental results on the task of opin- ion target identification show that RNNs, without using any hand-crafted features, outperform feature-rich CRF-based mod- els. Our framework is flexible, allows us to incorporate other linguistic features, and achieves results that rival the top perform- ing systems in SemEval-2014. 1 Introduction Fine-grained opinion mining involves identifying the opinion holder who expresses the opinion, de- tecting opinion expressions, measuring their inten- sity and sentiment, and identifying the target or aspect of the opinion (Wiebe et al., 2005). For ex- ample, in the sentence “John says, the hard disk is very noisy”, John, the opinion holder, expresses a very negative (i.e., sentiment with intensity) opin- ion towards the target “hard disk” using the opin- ionated expression “very noisy”. A number of NLP applications can benefit from fine-grained opinion mining including opinion summarization and opinion-oriented question answering. The tasks in fine-grained opinion mining can be regarded as either a token-level sequence labeling problem or as a semantic compositional task at the sequence (e.g., phrase) level. For example, iden- tifying opinion holders, opinion expressions and opinion targets can be formulated as a token-level sequence tagging problem, where the task is to The hard disk is very noisy O B-TARG I-TARG O O O O O O O B-EXPR I-EXPR Table 1: An example sentence annotated with BIO labels for opinion target (TARG tags) and for opin- ion expression (EXPR tags) extraction. label each word in a sentence using the conven- tional BIO tagging scheme. For example, Table 1 shows a sentence tagged with BIO scheme for opinion target (middle row) and for opinion ex- pression (bottom row) identification tasks. On the other hand, characterizing intensity and sentiment of an opinionated expression can be regarded as a semantic compositional problem, where the task is to aggregate vector representations of tokens in a meaningful way and later use them for sentiment classification (Socher et al., 2013). Conditional random fields (CRFs) (Lafferty et al., 2001) have been quite successful for different fine-grained opinion mining tasks, e.g., opinion expression extraction (Yang and Cardie, 2012). The state-of-the-art model for opinion target ex- traction is also based on a CRF (Pontiki et al., 2014). However, the success of CRFs depends heavily on the use of an …

Content truncated to 500 words