A Review on Sentiment Classification of Amazon Product Review dataset using NLP technique

Abstract

The objective of the review paper is to explore recent advancements in sentiment analysis of Amazon product reviews using Natural Language Processing (NLP) techniques. The paper aims to assess various methods and models used for classifying consumer sentiment as positive, negative, or neutral, emphasizing the role of deep learning algorithms and pre-trained word embedding’s. The methodology section discusses several state-of-the-art models, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory networks (LSTMs), and transformer-based models like BERT, which are utilized to improve sentiment classification accuracy. The paper also highlights the significance of embedding techniques such as Word2Vec, GloVe, and FastText, which capture the semantic meaning and contextual relationships in textual data. The results of various studies and experiments conducted in the field are presented, showing that deep learning models, particularly BERT and LSTMs, outperform traditional machine learning models in terms of accuracy, precision, and recall. The paper also discusses challenges in sentiment analysis, including handling noise, sarcasm, and multilingual reviews, and the importance of model refinement to address these issues. Finally, the review concludes by emphasizing the importance of NLP in understanding consumer sentiment and its practical applications for businesses in improving product offerings and customer engagement. This review offers a comprehensive overview of the techniques, challenges, and results in sentiment analysis of Amazon reviews, showcasing the transformative potential of NLP and deep learning technologies.