Product Recommendation System

Developed an advanced recommendation system leveraging Autoencoders to identify latent user preferences.

This approach improved recommendation accuracy by 20% through effective feature extraction and dimensionality reduction using PCA, enhancing user satisfaction.

The model was trained on a diverse dataset of Amazon product reviews, employing techniques like batch normalization and dropout to improve generalization and prevent overfitting.

The final system was deployed using Flask, providing real-time product recommendations.