Movie Genre Classifier (93% Accuracy)

Overview

A supervised learning project that classifies movie genres using textual metadata and TF-IDF features.

Problem

Manual genre classification is slow and inconsistent when datasets grow. The goal was to automate labeling with strong accuracy.

Solution

Built a preprocessing pipeline and trained a Random Forest model on TF-IDF vectors, then packaged results in a Streamlit demo for easy exploration.

Outcomes

• Achieved 93% accuracy on 5,000+ records.
• Delivered a working Streamlit demo for stakeholders.
• Led the final project presentation and results walkthrough.

My Contributions

• Performed data cleaning, preprocessing, and feature extraction.
• Trained and tuned a Random Forest classifier with TF-IDF vectors.
• Led the final presentation and delivered model insights.

Tech Stack

Features

• TF-IDF based text vectorization
• Random Forest classification pipeline
• Streamlit demo for interactive predictions

Challenges & Learnings

• Balancing precision and recall across multi-genre categories.
• Communicating model performance to non-technical audiences.