Enhancing Apple Product Review Sentiment Analysis using Machine Learning and Genetic Algorithm Optimization

No Thumbnail Available
Sugunan, Sruthy
Issue Date
MSc in Data Analytics
Dublin Business School
Items in eSource are protected by copyright. Previously published items are made available in accordance with the copyright policy of the publisher/copyright holder.
Using cutting-edge machine learning techniques along with Genetic Algorithm (GA) for optimization, this research delivers a thorough study of sentiment analysis of customer evaluations for Apple items. Understanding consumer attitudes can be essential in guiding product improvements and marketing tactics given the enormous amount of user-generated data. The goal of this research is to create a precise and effective model that can automatically categorize customer evaluations as either favourable, negative, or neutral. The study initially goes through a lengthy preprocessing stage to purge and appropriately prepare the data. The next step is to train three well-known classification algorithms to identify if sentiment is positive, negative, or neutral. The sentiment classification was done based on the polarity score and star-based method These methods are Logistic Regression, Support Vector Classification (SVC), and Naive Bayes. The best-performing model is determined through a thorough assessment procedure, and it is then further optimized by Genetic Algorithm utilizing the Tree based Pipeline Optimization Tool (TPOT). AUC, F1-score, accuracy, precision, recall, and other performance indicators are used to carefully examine the optimization's result. The comparison of the classifiers in polarity score-based sentiment analysis and the star-based analysis was done and it was found that the star-based analysis gave maximum accuracy 0.93 for SVM classifier with genetic programming optimization using TPOT. The project's findings offer insightful information about how consumers feel about Apple products and show how successful a coordinated strategy combining conventional ML algorithms with GA optimization can be. The outcomes not only aid in understanding consumer preferences and impressions but also present a novel technique that is transferable to various fields and goods. Throughout the project, legal standards for site scraping are observed, maintaining ethical integrity. Therefore, the study has significant ramifications for Apple's relentless pursuit of product development and client pleasure