Amazon-Reviews-Sentiment-Analysis

Project Description:

This project does the sentiment analysis of Amazon Reviews using two different approaches, and then comparing them.

• The dataset consists of 568454 records with 10 features.
• The main features include 'Score' (1-5) and 'Text' (on which the sentiment analysis is done).
• After completing the sentiment analysis on 'Text', the results are compared with 'Score' to check the correlation.
• The two models used for sentiment analysis are compared with respect to time and accuracy. This is done using Data visualization with pairplot.

Techniques used

Natural Language Toolkit

• Used NLTK for tokenization, tagging and parsing.

VADER Model

• Valence Aware Dictionary and sEntiment Reasoner
• The VADER model is faster, but does not consider the context of the sentence.

Roberta pretrained model

• This model is pretrained on ~58M tweets and finetuned for sentiment analysis.
• Transformer based model which accounts for the words but also the context related to other words.
• More confident predictions compared to the VADER model.
• More time consuming than the VADER model.