Deep clustering of small molecules at large-scale via variational autoencoder embedding and K-means

Getting Data

Visit: https://www.chemicalgenomicsoftb.com

The raw dataset that used in this study (47217 SMILES) is available there.

Generate The Features

RUN:
feature_generation.py

Data Pre-Processing

RUN:
Data preprocessing.py

Choose The best Number of Clusters

RUN:
Choose the best K.py

Training Models and Clustering algorithms

RUN:

• VAE & K-Means model&train.py
• AE & K-Means_model&train.py
• BIRCH.py

Getting The Results

RUN:
* K-Means & Internal clustering evaluations.py
* Visualize_results.py