Aspect based Sentiment Analysis in Hindi: Resource Creation and Evaluation

Aspect based Sentiment Analysis in Hindi: Resource Creation and Evaluation Md Shad Akhtar, Asif Ekbal and Pushpak Bhattacharyya Department of Computer Science and Engineering Indian Institute of Technology Patna Patna, India {shad.pcs15,asif,pb}@iitp.ac.in Abstract Due to the phenomenal growth of online product reviews, sentiment analysis (SA) has gained huge attention, for example, by online service providers. A number of benchmark datasets for a wide range of domains have been made available for sentiment analysis, especially in resource-rich languages. In this paper we assess the challenges of SA in Hindi by providing a benchmark setup, where we create an annotated dataset of high quality, build machine learning models for sentiment analysis in order to show the effective usage of the dataset, and finally make the resource available to the community for further advancement of research. The dataset comprises of Hindi product reviews crawled from various online sources. Each sentence of the review is annotated with aspect term and its associated sentiment. As classification algorithms we use Conditional Random Filed (CRF) and Support Vector Machine (SVM) for aspect term extraction and sentiment analysis, respectively. Evaluation results show the average F-measure of 41.07% for aspect term extraction and accuracy of 54.05% for sentiment classification. Keywords: Aspect based Sentiment Analysis, Sentiment Dataset, Indian Languages. 1. Introduction With the globalization of Internet, web generated contents are increasing at a tremendous pace. This huge amount of data has introduced several new challenges and opportuni- ties in the research communities. These days customers or users are relying heavily on other user’s opinion about a product or service before experiencing themselves. In or- der to get an unbiased opinion one has to extract and read all the reviews which is not an easy task to perform. Senti- ment Analysis (Pang and Lee, 2008) refers to the problem of automatically determining the polarity of sentiment/opinion expressed by user in a chunk of text or review. In gen- eral, polarity of a review belongs to one of the four possi- ble classes: positive, negative, neutral and conflict. Aspect based sentiment analysis (ABSA) is a fine-grained analysis of sentiments at the aspect or feature or attribute level. The term ‘aspect’ refers to an attribute or a component of the product/service that has been commented on in a review. Overall problem of aspect based sentiment analysis can be thought as a two-step process. The first step, i.e., aspect term extraction focuses on identifying various terms that denote aspects, and the second step, i.e. sentiment classi- fication deals with classifying the sentiments with respect to the aspect. A review sentence, therefore, may contain more than one aspect term and the sentiment associated with each. Such a fine-grained analysis provides greater insight to the sentiments expressed in the written reviews. In re- cent times, there have been a growing trend for sentiment analysis at the more fine-grained level, i.e. for aspect based sentiment analysis (ABSA). Some of the recent systems that have emerged are (Toh and Wang, 2014; Chernyshevich, 2014; Wagner et al., 2014; Castellucci …

Content truncated to 500 words