Text Mining for the Identification and Synopsization of Trusted Ratings

Online reviews have become increasingly valuable in the realm of purchasing and business decision-making. They not only contribute to brand loyalty but also aid customers in understanding product experiences. Unfortunately, the rise of fraudulent reviews, where authors are rewarded for generating fake evaluations to manipulate readers' perceptions, is a growing concern. This research paper aims to address this issue by developing a product review summarizer specifically designed to generate summaries based on genuine Amazon product reviews. The summarizer utilizes text analysis techniques to assess the linguistic characteristics of sentences and determine their value. The paper provides a comprehensive overview of existing and previous studies on sentiment analysis and text description. To enhance classification accuracy, a new hybrid classification system is proposed, combining multiple classification methods using arcing classifiers. The quality of these classifiers is thoroughly evaluated. In this study, supervised machine learning algorithms are compared with SVD dimensionality reduction, deep learning, and a text mining approach for summarization purposes. The model is built using a labeled Amazon review dataset. The key contribution of this paper lies in its innovative approach of generating text summaries for Amazon reviews while effectively filtering out fraudulent content.