Bio-Inspired Approach for Estimation of Parkinson’s Disease Using Augmented Feature Selection Model

The quality of life for millions of people worldwide is impacted by Parkinson's disease (PD), one of the most prevalent neurological diseases. The Unified Parkinson's Disease Rating Scale (UPDRS) is widely utilized to evaluate this degenerative neurological condition that impairs brain function. PD is a degenerative, long-term neurological condition that impairs movement. Medication can be started earlier with the aid of early detection. It can help the patient sustain a high quality of life for a longer period by considerably slowing down the disease's course. Researchers have recently underlined the need to look at different parts of the human brain to examine changes occurring in brain tissue and to gain a more thorough understanding of PD. Therefore, it is necessary to do precise diagnostics and treatment planning to detect the disease early. The submitted work proposes an augmented feature selection and estimation (AFSE) with a network model to forecast the course of PD utilizing the monitoring dataset's condensed input feature space. The new optimized architecture with a norm penalty (L2) tuned parameter receives the reduced input feature space and then uses it to analyze the prediction performance in PD progression. Several experiments are conducted to assess the performance, and the outcomes are compared to those of previously developed algorithms on the same dataset.