composer-classification

Music composer classification from MIDI files using machine learning and extracted musical features.

This repository contains a deep learning project for classifying classical music composers. It uses LSTM and CNN models to predict composers like Bach, Beethoven, Chopin, and Mozart, with an additional analysis covering 147 composers. The repo includes data, code, and documentation for replication and exploration.

Reproduces a segment-level Beethoven classifier (Galajda & Hua, MIPR 2024) and shows it is fooled by imitations spliced from real Beethoven bars, while a detector that reads bar-to-bar transitions is not.