Objective: Using machine learning techniques to automatically recommend suitable drugs for Parkinson’s Disease (PD) patients, thereby providing computer-aided prescriptions.

Background: PD is the second most common degenerative disease, of which the incidence is increasing year by year. PD seriously threatens the health of the elderly, and brings enormous economic pressure. However, not every doctor is specialized in the use of clinical drugs for Parkinson’s disease, which could lead to the early appearance of the patient’s movement complications. With the development of the artificial intelligence, which has already been successfully applied in medical diagnosis, automatic drug recommendation could be implemented to offer the patients better curative effect.

Methods: We evaluated the data (including drugs, motor symptoms and nonmotor symptoms) collected from 136 PD patients which took us two years for collection. Moreover, we present a novel model on these date via Learning lAtent Symptoms (PALAS), which could linked the observed symptoms with the known drug by our analysis.

Results: By Learning lAtent Symptoms (PALAS) methods, the novel model, which linked the observed symptoms with the known drug collected from 136PD patients, demonstrate the effectiveness for the PD drugs prescription task.

Conclusions: The Computer-Aided Prescription Model for Parkinson’s Disease could predict the suitable drugs for Parkinson’s Disease (PD) patients.

« Back to 2018 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/a-computer-aided-prescription-model-for-parkinsons-disease-could-predict-the-suitable-drugs-for-parkinsons-disease-pd-patients/