Product Reviews Based Sentimental Analysis by Feature Extraction and Classification Using Machine Learning Algorithm

Sentiment analysis and opinion mining are two of the most popular domains for evaluating and extracting insights from text data from various sources, like Facebook, Twitter, Amazon, and so on. It is essential for giving companies the ability to actively work on refining their company plans and thoroughly understand what customers have to say about their offerings. This research proposes novel technique in product review based sentimental analysis by online reviews data feature extraction with classification using machine learning model. here the input has been collected as online review data and processed for noise removal and normalization. Then, this data features are extracted and classified using graph recurrent component analysis with support Bayesian fuzzy neural network (GRCA-SBFNN) for detection of the user sentiment for the reviewed product. the experimental analysis has been carried out for various online product review dataset in terms of training accuracy, average precision, area under the curve (AUC), F-1 score and recall. The proposed technique Training accuracy 98%, Average precision 97%, AUC 91%,Recall 96%, F-1 SCORE 92%.