Rough Set Data Mining Algorithms and Pursuit Eye Movement Measurements Help to Predict Symptom Development in Parkinson’s Disease

Abstract

This article presents research on pursuit eye movements tests conducted on patients with Parkinson’s disease in various stages of disease and phases of treatment. The aim of described experiment was to develop algorithms allowing for measurements of parameters of pursuit eye movement in order to reference calculated results to the previously collected neurological data of patients. An additional objective of the experiment was to develop an example of data-mining procedure, allowing for classification of neurological symptoms based on oculometric measurements. Definition of such correlation enables assignment of particular patient to a given neurological group on the base of parameters values of pursuit eye movements. By using created decision table, we have achieved good results of prediction of the neurological parameter UPDRS, with total accuracy of 93.3% and total coverage of 60%. This allows for better evaluation of stage of the disease and its progression. It also might provide additional tool in determining efficacy of different disease treatments.