Data Mining Applied to Diagnose Diseases Caused by Lymphotropic Virus: a Performance Analysis

This paper proposes a new methodology to diagnose the rheumatology manifestations and HTLV-I-Associated Myelopathy/Tropical Spastic Paraparesis, or HAM/TSP, in patients who have Lymphotropic virus of T cells in Humans or HTLV of type I and II. Computational intelligence algorithms are used to classify HTLV patient carriers with or without the presence of rheumatology manifestations and of HAM / TSP. A benchmarking is performed among artificial neural intelligence, naïve bayes, Bayesian networks and decision tree to evaluate the most suitable technique for solving this application issue. The obtained results demonstrate the potential of the methodology on the helping non-specialist doctors to classify the patient with the disease suspicion.