The Impact of Spam Reviews on Feature-based Sentiment Analysis

With the explosive growth of online social media, people can convey their experience and share with others on a common platform. People's opinions and experience became the most important source of information in the process of sentiment analysis for making decisions. The performance of opinion mining depends on the availability of trustworthy opinions for sentiment analysis. Unfortunately, spammers write spam reviews which can be positive or negative opinions in order to promote their services or damage the reputation of their competitors' services. Spam reviews may mislead potential customers and affect their experience and influence their ideas. These spam reviews must be identified and removed to avoid possible deceitful customers. The main objective of this paper is to present an enhanced feature-based sentiment analysis algorithm that improves the performance of sentiment classification. The proposed algorithm is developed to assign accurate sentiment score to each feature in social reviews by considering spam reviews detection. The proposed algorithm also examines the effect of three different feature extraction methods on the performance of sentiment classification. Finally, the results indicate that the proposed algorithm achieves an accuracy of about 79.56% in classifying opinions.