IITP: Supervised Machine Learning for Aspect based Sentiment Analysis

Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014), pages 319–323, Dublin, Ireland, August 23-24, 2014. IITP:Supervised Machine Learning for Aspect based Sentiment Analysis Deepak Kumar Gupta Indian Institute of Technology Patna Patna, India deepak.mtmc13@iitp.ac.in Asif Ekbal Indian Institute of Technology Patna Patna, India asif@iitp.ac.in Abstract The shared task on Aspect based Senti- ment Analysis primarily focuses on mining relevant information from the thousands of online reviews available for a popular product or service. In this paper we re- port our works on aspect term extraction and sentiment classification with respect to our participation in the SemEval-2014 shared task. The aspect term extraction method is based on supervised learning algorithm, where we use different classi- fiers, and finally combine their outputs us- ing a majority voting technique. For senti- ment classification we use Random Forest classifier. Our system for aspect term ex- traction shows the F-scores of 72.13% and 62.84% for the restaurants and laptops re- views, respectively. Due to some techni- cal problems our submission on sentiment classification was not evaluated. However we evaluate the submitted system with the same evaluation metrics, and it shows the accuracies of 67.37% and 67.07% for the restaurants and laptops reviews, respec- tively. 1 Introduction Nowadays user review is one of the means to drive the sales of products or services. There is a grow- ing trend among the customers who look at the on- line reviews of products or services before taking a final decision. In sentiment analysis and opinion mining, aspect extraction aims to extract entity as- pects or features on which opinions have been ex- pressed (Hu and Liu, 2004; Liu, 2012). An aspect is an attribute or component of the product that This work is licensed under a Creative Commons At- tribution 4.0 International License. Page numbers and pro- ceedings footer are added by the organizers. License details: http://creativecommons.org/licenses/by/4.0/ has been commented on in a review. For exam- ple:“Dell Laptop has very good battery life and click pads”. Here aspect terms are battery life and click pads. Sentiment analysis is the task of iden- tifying the polarity (positive, negative or neutral) of review. Aspect terms can influence sentiment polarity within a single domain. As an example, for the restaurant domain cheap is usually posi- tive with respect to food, but it denotes a negative polarity when discussing the decor or ambiance (Brody and Elhadad, 2010). A key task of aspect based sentiment analysis is to extract aspects of entities and determine the sentiment corresponding to aspect terms that have been commented in review document. In recent times there has been huge interest to identify as- pects and sentiments simultaneously. The method proposed in (Hu and Liu, 2004) is based on infor- mation extraction (IE) approach that identifies fre- quently occurring noun phrases using association mining. Some other works include the methods, viz those that define aspect terms using a manually specified subset of the Wikipedia category (Fahrni and Klenner, 2008) hierarchy, unsupervised clus- tering technique (Popescu and …

Content truncated to 500 words