Music Genre Classification
Python, Fall 2019
This project classifies music genres using more human intuitive feature such as those found in sheet music, including pitch, rhythm, tempo, and instruments. In order to extract relevant patterns in temporal and spatial information from these features, long-short term memory (LSTM) re-current neural networks were used for the classification task. Despite a messy dataset and minimal features, we were able to classify 5 genres with an accuracy of approximately 63%.