Combination of Trees for Guillain-Barré Subtype Classification.

Guillain-Barre Syndrome (GBS) is an autoimmune neurological disorder characterized by a fast evolution. Complications of this disorder vary among the different subtypes. In this study, we use a real dataset that contains clinical, serological, and nerve conduction test data obtained from 129 GBS patients. We apply three different decision tree classifiers: C4.5, C5.0 and random forest to predict GBS subtypes in two classification scenarios: four subtype classification and One vs All (OVA) classification. We evaluate performance under train-test scenario. Experimental results showed comparable performance among all classifiers, although C5.0 slightly outperformed both C4.5 and random forest. Further experiments are being conducted. This is an ongoing research project.