Aspect Based Online Sentiment Analysis Product Review and Feature Using Machine Learning

Abstract

Today, people exchange their thoughts through online web forums, blogs, and different social media platforms. In online shopping, they provide reviews and opinions on various products, brands, and services. Their thoughts not only influence the purchasing decisions of consumers but also contribute to improving product quality based on their requirements and identifying specific problems with products to find excellent solutions. The present system focuses on the peer-reviewed review model (user-generated reviews) and global qualification, such as ratings. It attempts to classify the semantic aspects and emotions at the time aspect level from the data to understand the overall sentiment conveyed in the reviews. SJASM represents each review document in the format of opinion pairs, simulating the appearance terms and corresponding opinion words while considering hidden aspects and sentiment detection. The current system is designed as a recommendation system using Natural Language Processing (NLP) techniques to analyze reviews and automatically classify them using Naïve Bayes Classification. Additionally, we have extracted the product characteristics from the reviews, allowing administrators to analyze opinion pairs to identify defects in finished products, thus potentially increasing the market for those products in the future. This system aims to extract product aspects and corresponding opinions from consumer ratings on the internet. Different machine learning algorithms, including Naïve Bayes, are discussed for sentiment classification, and performance metrics such as precision, recall, F-score, and accuracy are used to evaluate classifier performance.