NLP - Urdu Sentiment Analysis on Twitter Data (2020)

This repository contains code for performing sentiment analysis on Urdu tweets to understand the perception of Pakistani citizens regarding their ARMY and Judiciary using Natural Language Processing (NLP) techniques. The dataset used for this analysis was generated through web scraping from Twitter and some newspaper sites.

Tools Used

• UrduHack: A NLP library for Urdu language processing.

Machine Learning Models and Vectorizers Used

Vectorizers

• TF-IDF Vectorizer
• Word2Vec
• Doc2Vec

Models

• Random Forest Classifier
• Logistic Regression
• Support Vector Machine (SVM) Classifier
• Multinomial Naive Bayes
• Stochastic Gradient Descent (SGD) Classifier
• Recurrent Neural Network (RNN): Utilized for deep learning-based sentiment analysis.

Dependencies

• pandas
• numpy
• gensim
• scikit-learn
• keras
• tensorflow