LDA topic modeling of raw audio data for music suggestions
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore
README.md
analysis.py
datasets.py
mp3_to_wav.py
mp3test.py
test.py
train.py

README.md

LDA-Music

Using LDA topic modeling to generate music suggestions using a model trained on audio frequency data.

Dataset

Currently using the GTZAN Genre Collection of 1000 song samples, each 30 seconds long. The tracks are 16/22050 mono files in the .au format. (http://marsyasweb.appspot.com/download/data_sets/)

This data set came out of work from the following publication.

"Musical genre classification of audio signals" by G. Tzanetakis and P. Cook in IEEE Transactions on Audio and Speech Processing 2002.

Dependencies

Written for Python 2.7.13