Unsupervised Learning for Hydrogen Breath Tests

Hydrogen breath tests are a well-established method to help diagnose functional intestinal disorders such as carbohydrate malabsorption or small bacterial overgrowth. In this work we apply unsupervised machine learning techniques analyze hydrogen test datasets. We propose that uses 26 internal cluster validation measures determine suitable number of clusters. an induced external step use predefined categorization proposed by medical expert. The results indicate the majority considered indexes was not able produce reasonable clustering. Considering performed expert, novel shape-based obtained highest measure in terms adjusted rand index. clusterings constitute basis supervised is part our ongoing research.