Interpretable-Sequence-Clustering-Tree

Source Code for Interpretable sequence clustering https://www.sciencedirect.com/science/article/pii/S0020025524013677

Running

We added the tree structure visualization, to use it on linux please first

sudo apt-get install graphviz
export PATH=$PATH:/usr/local/bin
source ~/.bashrc

We recommend using Pypy, which may give exponential speedups on larger datasets

Usage

Random Projection on Python version

pip install ISCT

from ISCT import ISCT
sequences = [
    ['a', 'g', 't', 't', 'c'], 
    ['a', 't', 'g', 'g', 't', 'u', 't'], 
    ['a', 'c', 't', 'u', 'u', 'a', 'a'], 
    ['a', 'c', 'a', 'u', 'a', 't', 'c', 't'], 
    ['a', 'g', 'g', 'c', 'a', 'a', 'c'], 
    ['a', 'c', 'g', 'g', 'c', 'c', 'a', 'a']]
isct = ISCT(num_clusters = 3, visulization_name = 'test', min_sample=2) # if visulization_name is provided, it generates the "test.pdf".
y_pred = isct.fit(sequences)
print(y_pred)

Dependencies

• Python 3.9.16 (Pypy)
• NumPy 1.24.3
• Scikit-learn 1.2.2
• Pandas 2.0.1
• Prefixspan 0.5.2

Visualization

ISCT could provide you with a highly concise and short clustering tree, taking poineer as example: