Sentiment Analysis on IMDb Reviews Using NLP

Abstract

Sentiment analysis technologies enable organizations to analyze consumer sentiments in online feedback about products, brands, and services. It enables us to determine the mood of the audience. Using the text analysis methods, sentiment analysis measures and classifies the emotions (whether it is positive, negative, and neutral) represented in the text data. IMDb contains information about movies, TV shows, web series, video games, and other media, such as cast, crew, storyline summaries, ratings, and fan and critic reviews. Sentiment Analysis (SA) uses Natural Language Processing (NLP) to provide a common machine learning (ML) solution to this problem. In the proposed study, we performed sentiment analysis on the IMDb movie reviews dataset from Kaggle’s Bag of Words meets Bag of Popcorn to demonstrate how substantial insights can be extracted from a massive quantity of textual data obtained from the internet. The Random Forest (RF) model, a standard machine learning method, is used to obtain these insights. In addition, the Python modules Beautiful Soup and Count Vectorizer are used to translate a given text into a vector depending on the frequency (count) of each word in the full text.