Parkinson’s Disease Detection using Machine Learning: Current Trends and Future Perspectives

Abstract

To begin, Parkinson defines Parkinson’s disease as a neurologic illness that affects the central nervous system, causing sufferers to have trouble communicating, walking, and trembling throughout the movements. Parkinson’s disease patients often have low-volume noise with a monotone quality; this technique investigates the classification of audio signals feature dataset to diagnose Parkinson’s disease (PD); the classifiers utilised during this system are from Machine Learning.Our model typically makes use of provision regression and XGboost classifiers, other than the audio feature dataset from the UCI dataset repository. The system got a far superior outcome in predicting whether the palladium patient is healthy or not, with XGBoost providing a height accuracy of 96% and a Matthewsparametric statistic (MCC) of 89%.
Parkinson’s disease is a neurological condition that affects millions of individuals throughout the world. 60% of people over the age of 50 have Parkinson’s disease (PD). Patients with Parkinson’s disease have difficulty communicating and moving, making it difficult for them to go to treatment and monitoring visits. Early detection of Parkinson’s disease allows for treatment, allowing people to live normal lives. The world’s ageing population emphasises the importance of detecting Parkinson’s disease early, remotely, and properly. Machine learning approaches have showed considerable promise in the early identification and diagnosis of Parkinson’s disease in recent years.We present a unique strategy for detectingParkinson’s illness utilising machine learning techniques and the Xception architecture in this research.
We concentrate on detecting Parkinson’s disease using spiral and wave drawings, which are routinely employed in clinical practise as bite of the diagnostic process.We collected spiral and wave drawings from people with and without Parkinson’s disease. We preprocessed the data and proficient the models with ML the Xception framework.
Our models performed admirably, with a training accuracy of 95.34% and a validation accuracy of 93.00% for Parkinson’s disease detection from spiral drawings, and a training accuracy of 93.34% and a validation accuracy of 86.00% for Parkinson’s disease detection from wave drawings. Our findings show that machine learning and the Xception architecture have the ability to detect and diagnose Parkinson’s disease early. Our strategy has the potential to increase the accuracy and speed of Parkinson’s disease diagnosis, resulting in improved patient outcomes and quality of life.