Sentiment Analysis & Movie Recommendation

Combined sentiment analysis with clustering for more relevant movie recommendations using IMDB reviews.

Overview

Developed an NLP-driven system that processes 10,000+ IMDB reviews to classify sentiment and improve recommendation quality. By combining traditional text processing with clustering, the system produced more personalized movie suggestions.

Approach

  • TF-IDF Vectorization to convert text into features.
  • K-Means Clustering to group reviews by sentiment patterns.
  • PCA for dimensionality reduction and faster runtime.
  • Hyperparameter Tuning for accuracy optimization.

Results

  • Improved sentiment classification precision and reduced runtime by 50%.
  • Recommendations better aligned with user preferences.

Skills demonstrated: NLP (TF-IDF), clustering, dimensionality reduction, model tuning.