Amazon Product Reviews Sentiment Analysis Using ALBERT and Boosting Algorithms

Computer Engineering

Sentiment analysis plays a crucial role in understanding customer feedback and improving product quality. This project presents a machine learning approach that focuses on classifying customer product reviews based on sentiment using advanced techniques and boosting algorithms. The aim is to streamline the process of assessing customer opinions, enabling product owners to enhance their offerings efficiently, rather than manually analyzing each review. The project leverages the Amazon Product Reviews dataset from Kaggle.com to train and test various machine learning models, with a primary focus on four key algorithms: BiLSTM, CatBoost, XGBoost, and LightGBM. Each of these algorithms demonstrates impressive accuracy in sentiment classification, with BiLSTM achieving 93.17% and LightGBM reaching 90.38%, showcasing their proficiency in capturing the nuances of customer sentiments. To further enhance the accuracy of sentiment analysis, two distinct modules, Count Vectorizer and Tf-IDF Vectorizer, are incorporated to preprocess and represent the text dataset. This project systematically compares the performance of these modules, shedding light on the most effective approach for training the models.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.