Implicit Feature Detection via a Constrained Topic Model and SVM

Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 903–907, Seattle, Washington, USA, 18-21 October 2013. c⃝2013 Association for Computational Linguistics Implicit Feature Detection via a Constrained Topic Model and SVM Wei Wang ∗, Hua Xu ∗and Xiaoqiu Huang † ∗State Key Laboratory of Intelligent Technology and Systems, Tsinghua National Laboratory for Information Science and Technology, Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China †Beijing University of Posts and Telecommunications, Beijing 100876, China ww880412@gmail.com, xuhua@tsinghua.edu.cn, alexalexhxqhxq@gmail.com Abstract Implicit feature detection, also known as im- plicit feature identification, is an essential as- pect of feature-specific opinion mining but previous works have often ignored it. We think, based on the explicit sentences, sever- al Support Vector Machine (SVM) classifier- s can be established to do this task. Never- theless, we believe it is possible to do bet- ter by using a constrained topic model instead of traditional attribute selection methods. Ex- periments show that this method outperforms the traditional attribute selection methods by a large margin and the detection task can be completed better. 1 Introduction Feature-specific opinion mining has been well de- fined by Ding and Liu(2008). Example 1 is a cell phone review in which two features are mentioned. Example 1 This cell phone is fashion in appear- ance, and it is also very cheap. If a feature appears in a review directly, it is called an explicit feature. If a feature is only implied, it is called an implicit feature. In Example 1, appearance is an explicit feature while price is an implicit fea- ture, which is implied by cheap. Furthermore, an ex- plicit sentence is defined as a sentence containing at least one explicit feature, and an implicit sentence is the sentence only containing implicit features. Thus, the first sentence is an explicit sentence, while the second is an implicit one. This paper proposes an approach for implicit fea- ture detection based on SVM and Topic Model(TM). The Topic Model, which incorporated into con- straints based on the pre-defined product feature, is established to extract the training attributes for SVM. In the end, several SVM classifiers are con- structed to train the selected attributes and utilized to detect the implicit features. 2 Related Work The definition of implicit feature comes from Liu et al. (2005)’s work. Su et al. (2006) used Point- wise Mutual Information (PMI) based semantic as- sociation analysis to identify implicit features, but no quantitative experimental results were provided. Hai et al. (2011) used co-occurrence association rule mining to identify implicit features. However, they only dealt with opinion words and neglected the facts. Therefore, in this paper, both the opinions and facts will be taken into account. Blei et al. (2003) proposed the original LDA us- ing EM estimation. Griffiths and Steyvers (2004) applied Gibbs sampling to estimate LDA’s parame- ters. Since the inception of these works, many vari- ations have been proposed. For example, LDA has previously been used to construct attributes for clas- sification; it often …

Content truncated to 500 words