Music Mood Detection Using Machine Learning

Overview

Developed a machine learning model to analyze and predict emotional qualities in music, focusing on valence (positive/negative emotions) and arousal (intensity). This project combined audio signal processing with machine learning techniques to explore Music Emotion Recognition (MER).

Course: Music Informatics
Date: August–October 2024
Collaborators: Mingcheng Kou, Daichi Taguchi, Haoyun Zhou

Key Skills and Tools

  • Implemented machine learning models for regression and classification tasks.

  • Processed and analyzed large-scale audio datasets to extract meaningful features.

  • Tools used: Python, scikit, XGBoost, Pandas, NumPy.

  • Gained hands-on experience in model optimization, feature selection, and evaluation metrics.