Abstract
Due to increased usage of technology (social media, online marketing, internet services) over the years, Sentiment Analysis which is analysis done to understand the opinion ex-pressed in a piece of text, has become a hot and trending topic in the world today. Sentiment Analysis opens the door to a plethora of intriguing applications in almost every possible domain, from politics to social media rant on trending topics, education, movies, product and service reviews. A thriving aspect of Sentiment Analysis is the sentiment polarity in customer reviews; companies receive criticism and commendation from customers to understand their views on different products and determine which products are more favourable than others. This project attempts to build two deep learning models: Convolutional Neural Network (CNN) & Long Short-Term Memory (LSTM) to automatically detect and classify sentiment polarity in Amazon Electronic review dataset. The raw text is processed into their respective word vector representation using GloVe Embeddings. Accuracy, Precision, Recall, and F1 Score are used to assess the selected models. Both
baseline models obtain 93% Accuracy. The results demonstrate that the models are able to accurately classify the reviews.
