Feature based Opinion Mining using Convolutional Neural Networks

Abstract

Opinion mining refers to the use of natural language processing, text analysis to identify and extract user’s opinion through their reviews. Now a day’s people before purchasing any product they are reading reviews given by earlier customers, based on these reviews they are making decision to purchase the product. Opinion mining is the automatic process of determining the polarity of the product review. To get the exact polarity Feature based opinion mining works prominently than Document and Sentence level. Feature Based Opinion Mining (FBOM) is the task of classifying the sentiment of a specific feature in a text. Because a single text usually has multiple aspects which are expressed independently, FBOM is an important task for in-depth opinion mining. Opinion mining is the improvement of granularity at feature level, representing two distinct aims: aspect extraction and sentiment classification of product and sentiment classification of target-dependent tweets. However, it is problematic to depend solely on attention mechanisms to achieve this, because most sentiment expressions such as ‘nice’ and ‘bad’ are two general to be aligned with a proper feature event through an attention mechanism. To full fill this task, this paper proposes a convolutional neural network (CNN) which is a deep learning model based feature level sentiment classification model, which consists of two CNNs. The first CNN extracts the positional information of feature expressions for a target feature. The second CNN classifies the sentiment of the target feature in text.