IMDb Sentiment Analysis

Streamlit Deployment Link

IMDb Sentiment Analysis

Project Details

Introduction

IMDb is an online database of information related to entertainment industry, it contains various data related to movies, television series, podcasts and several other form of media.

Objective

The objective of this project is to develop a sentiment analysis model for IMDb movie reviews. The problem is to analyze and classify a review or set of reviews into positive or negative while also considering the star rating.

Dataset Description

• The dataset was created by scraping data using Selenium and the IMDb ID of movies.
• Review title, star rating and actual review were scraped for each review of the movie.
• The final size of dataset was 111555 independent entries of reviews and respective user rating.

Python Libraries Used

Streamlit
Beautiful Soup 
Pandas
Plotly
Requests 
Scikit-learn

Models trained

1. KNN Classifier Algorithm
   • Accuracy: 73%
2. Multinomial Naïve-Bayes Algorithm
   • Accuracy: 76%
3. Logistic Regression Model
   • Accuracy: 87%
4. Decision Tree
   • Accuracy: 75%