Opinion Mining on Amazon Musical Product Reviews using Supervised Machine Learning Techniques

• Published in: 2023 11th International Conference on Internet of Everything, Microwave Engineering, Communication and Networks (IEMECON) pp. 1 - 6
• Main Authors: Kumar, Aryan, Jain, Tarun, Tiwari, Priyesh, Sharma, Rakesh
• Format: Conference Proceeding
• Language: English
• Published: IEEE 10.02.2023
• Subjects: ANN; Decision making; Decision Tree; Deep learning; K-Nearest Neighbour; Logistic Regression; Machine learning algorithms; Music; Naïve Bayes; Random Forest; RNN; Sentiment analysis; Social networking (online); Support vector machines; SVM (Support Vector Machine); TF-IDF
• DOI: 10.1109/IEMECON56962.2023.10092288

People these days express their sentiments about a specific product through social media and networking websites. Sentiment Analysis or Opinion Mining is the analysis of such sentiments from texts, which uses natural language processing. Opinion mining on Amazon musical product reviews identifies these sentiments by classifying them as a positive or a neutral or a negative polarity. This will also play a vital role in the decision making and recommendation of products by understanding the polarity of the reviews. In this proposed work, the Machine Learning feature extractions (TF-IDF and Bag of Words) give a much better accuracy when tested against 6 classifying algorithms; Naïve Bayes, Logistic Regression, Decision Tree, K-Nearest Neighbour, Random-Forest and SVM, and 2 neural networking algorithms; Artificial Neural Network and Recurrent Neural Network. The proposed work will help in distinguishing between positive, neutral and negative reviews accurately. Therefore, a large dataset of more than 10000 reviews, with an average of 300 words per review, is used to serve as a stronger model. During implementation, a maximum accuracy of 97.7% was implemented when tested using the SVM algorithm, and a minimum accuracy of 81.3% was implemented when tested using the Decision-Tree algorithm.