/Sentiment-Analysis-of-IMDB-Film-Reviews

Primary LanguageJupyter NotebookMIT LicenseMIT

Sentiment-Analysis-of-IMDB-Film-Reviews

This project was done as a part of computational intelligence course.

Workflow

1. Text data pre-processing 

a. Removing Punctuation and HTML Tags
b. Converting every letter 
c. Converting every number to corresponding text(i.e. '5' is converted to 'five') 
d. Removing Stop Words

2. Normalizing Text Data

a. Stemming
b. Lemmatization

3. Data Representation

a. Vectorization using Binary Representation
b. Vectorization using Word Count
c. Vectorization using TF-IDF

4. Data Classification

a. Logistic Regression Model
b. SVM Classifier

By Experimenting various model parameters, Finally broke the 90% accuracy mark!