Neurogenerative Disorder Identification From Speech and Hand Drawn Inputs Using Machine Learning Techniques

Abstract

Parkinson's disease symptoms mainly occur at the age of 60 years or older. There is a need for more accurate methods and objectives for early detection, which doctors can use in all cases to predict Parkinson's disease. The proposed hybrid approach analyses the patient's voice and hand-drawing data. Combining both results, the doctor may conclude the normal or abnormal conditions in patients and prescribe medication based on the category. The dimensions from the spiral drawings are obtained and analyzed using a machine learning algorithm; then, they are compared with trained data sets to get the required results. If both the voice and spiral drawing data state the presence of Parkinson's disease, then the disease is confirmed. If both the voice and spiral drawing state the absence of Parkinson's disease; the patient is without this disease. Finally, a linear regression machine learning algorithm is used to identify the presence of Parkinson's disease.