An experiment with a variety of clustering (and clustering-like) techniques to reduce noise on a speech audio recording.
Highlighted results could be found on this page.
- Add kmeans clustering
- Add hierarchical clustering
- Try to implement Word2vec
- Find more combinations between clustering algorithms and ICA
- Try to chain ICA results (use output as an input for the next run)
- Clean the code (it looks like a mess..)
Open-source python3 packages: