Music Genre Detection Using Audio Streams: A Machine Learning Approach

In the current digital age, the proliferation of music libraries and streaming services has led to an overwhelming abundance of musical content. Automated music genre detection is crucial for enhancing user experience through personalized recommendations, improved search functionalities, and dynamic playlist generation. Traditional genre classification methods re- liant on metadata are becoming impractical due to the vast volume of new music. Machine learning and artificial intelligence present promising solutions by leveraging advanced algorithms to discern patterns in audio files corresponding to different genres. This paper explores usage machine learning techniques on the GTZAN dataset for music genre detection. Models like Naive Bayes, SVM, Decision Trees, Random Forest, and Neural Networks are evaluated, Maximum accuracy has been achieved with CatBoost classifier (0.90524) and XGBoost Classifier (0.91058). Data preprocessing and PCA demonstrate significant improvements in accuracy. This study contributes insights into effective genre classification methods, benefiting listeners and music industry professionals.