Finding Sentiment Analysis for Amazon Product...

Finding Sentiment Analysis for Amazon Product Reviews using Machine Learning Algorithms

Author Name : Chandrashekar D K, Aditi Bansal, Akanksha, Ruchi Kumari, Udaya Shankar K R

ABSTRACT

One of the fields of study is sentiment analysis in computer science field which has the fastest pace of growth, making it challenging to catch pace with all the advancements in this domain. We use sentiment analysis, opinion mining, and text mining to offer user feedback evaluations on products that have altered public perceptions of a certain brand of goods. Online product reviews that were gathered from numerous websites served as the study's data source. We contrasted the responses that were obtained in terms of sentiment. Naive Bayes and Decision Tree were utilized for the segmentation of reviews among other classification models. A subtype of data filtering systems called recommendation engines aims to forecast the "ranking" or "preference" that a user would offer to an entity. By studying consumer decisions, one locates data designs in the informative index and generates outcomes that co-identify with customer needs. Machine learning algorithms are used in the classification process known as sentiment analysis to segregate a message into positive or negative regarding a certain topic. Three algorithms' performances—Multinomial Naive Bayes (MNB), Linear Support Vector Machine (LSVM), and Long Short-Term Memory. The best performance was achieved using the LSTM. The LSTM model was then tested on the additional 3.94 million ratings from the Kaggle database as well as a fresh dataset that was collected from Amazon and contained reviews on goods belonging to various categories. The best outcomes for the categorization were evaluations of furniture products. In summary, LSTM networks are excellent for categorizing the sentiment in product reviews, and the findings are quite consistent across categories. If the categorization is still valid when more than two classes are included, additional research is required.

Keywords: Amazon, Machine learning, Reviews, Sentiment Analysis.