An effective EBRBFSVM method for hybrid analysis of sentiments for the perspective of customer review summarization

Abstract

Corporate management has undergone a substantial rethinking recently and is now more than ever cantered on customer-oriented ideas. In truth, client connection administration technologies and procedures are becoming larger and greater common and essential for overcoming today’s company difficulties. To solve this issue, proposed work hybrid customer review summarization (CRS) and sentiment analysis (SA). CSR model would thus be ideal, since it can display the summarized data and give organizations valuable insight into the motivations behind consumers’ decisions and behavior. This article suggests SA from the standpoint of an efficient CRS. Pre-processing, feature extraction, and review categorization are some of the procedures involved in the task. Using natural language processing (NLP) and a variety of pre-processing methods, the pre-processing stage eliminates unnecessary information from text evaluations. It is suggested to create a bespoke feature vector for every client evaluation using a hybrid technique made up of aspect-related features and review-related features for effective feature extraction. Ensemble bootstrap-based radial basis function with support vector machines (EBRBFSVM) used for implementation. The testing findings demonstrate that the suggested EBRBFSVM completed the SA effectively and surpassed the currently available state-of-the-art approaches. Accuracy, F1-score, precision, and recall are used to compare the performances.