Clustering Optimization

An optimization method for clustering hyperparameters selection using internal validity indexes.

Requirements:

Docker Container Ubuntu 18.04

• Python 3

• Skopt

• Sklearn

Container

Create via Dockerfile:

docker build -t clustering_opt .

Clone via DockerHub:

docker pull marcosspalenza/clustering_opt

Usage

Requires input and output directories volumes as following.

docker run -v IN_DIR:/data/input/ -v OUT_DIR:/data/output/ clustering_opt:latest main_clustering.py DATASET

Replace IN_DIR and OUT_DIR with your data path and DATASET with your database name.

Also, the help options provide detailed descriptions:

docker run clustering_opt:latest main_clustering.py --help

---

Reference

Spalenza, M. A., Pirovani, J. P. C., and de Oliveira, E. Structures Discovering for Optimizing External Clustering Validation Metrics. In Proceedings of the 19th International Conference on Intelligent Systems Designand Applications, volume 19 ofISDA 2019, pages 150–161, Auburn (WA),USA. Springer International Publishing.