This project leverages machine learning to accurately predict song genres, similar to the technology used by popular music platforms like Spotify, Apple Music, and YouTube Music. Whether it's calm music for studying, energetic tunes for exercising, or music for everyday activities, this tool helps in categorizing and recommending music efficiently. The project was completed in the Spring of 2024 under CS 4641 at the Georgia Institute of Technology.
- Random Forest: Uses a supervised learning algorithm for genre prediction by constructing multiple decision trees and averaging their outputs for robust classification.
- Bayesian Networks: Evaluates song novelty and audio content for accurate classification.
- Neural Networks with Filtering: Combines neural networks and advanced filtering techniques for enhanced genre prediction accuracy.
Music platforms have explored various algorithms to enhance user experience by providing accurate music recommendations. Our project dives into:
- Random Forest: Training methodologies and application in music genre prediction based on a supervised set.
- Bayesian Networks: Assessment of song novelty and audio content.
- Neural Networks: Integration with filtering techniques for improved results.
- Python: Core programming language for implementing machine learning models.
- TensorFlow / PyTorch: Libraries for building and training neural networks.
- Scikit-Learn: Traditional machine learning algorithms and evaluation metrics.
The following technologies were used for the compilation of the project:
- Python 3.6+
- TensorFlow / PyTorch
- Spotify API
- Scikit-Learn
- Jupyter
Our model demonstrates a high accuracy rate in predicting song genres, making it a viable solution for integration into music recommendation systems.
- Supervised Learning
- Bayesian Networks for Song Novelty and Audio Content Evaluation
- Neural Networks with Filtering Techniques
We welcome contributions to improve our project. Feel free to fork the repository and create pull requests.
This project is licensed under the MIT License.
For more information, please contact pthakur31@gatech.edu.
We invite you to explore our project in more detail and contribute to its development!