A neurodegenerative condition that largely affects the central nervous system is Parkinson's disease, and subtle changes in the early stages of this disease make accurate diagnosis challenging. 'Bradykinesia', or sluggish movements, are among the disease's typical symptoms. The disease's symptoms start to show up in middle age, and as people age, the severity of the condition worsens. A speech issue is one of the first indications of Parkinson's disease. In this work, it is suggested that employing supervised classification algorithms for the subjective disease categorization, such as support vector machines (SVM), artificial neural networks (ANN), Naive Bayes (NB), etc. could be successful. The proposed method is compared with previously used Parkinson's disease diagnosis methods and well-known classifiers. The experimental results show that ANN is better than other supervised algorithms with the highest accuracy. The proposed work provides equivalent and superior outcomes.
Top1. Introduction
Parkinson's disease (PD) is a neurological problem that strikes millions of human beings worldwide. It is a long-term and persistent condition. In the brain's basement nigra, the depletion of dopaminergic neurons leads to many kinds of movement and non-motor signs and symptoms. (Federica, 2021) (Changqin, 2021). For prompt intervention, better therapeutic results, and patient quality of life improvement, early and correct PD diagnosis is essential. The early detection of the disease using conventional diagnosis approaches, which mostly rely on clinical evaluations and pricy imaging tools, is frequently difficult. Artificial intelligence (AI) and machine learning (ML) advancements in recent years have shown considerable promise in a variety of medical applications, including disease diagnosis. Identification of distinctive biomarkers and risk factors linked to PD is made possible by the capacity of machine learning (ML) systems to learn patterns and connections from enormous volumes of data. The purpose of this chapter is give an in-depth review of the current trends in using machine learning techniques for Parkinson's disease detection and to examine the potential for this cutting-edge method to completely transform the PD diagnostics industry. Recent studies have revealed that people with PD have unique voice traits and speaking patterns that set them apart from healthy people. Researchers are interested in these specific voice characteristics as potential biomarkers for an early diagnosis of Parkinson's disease (PD) (Mittapalli, 2023). These traits include diminished speech intensity, altered prosody, and vocal tremors. In this study, we investigate a novel method for Parkinson's disease early diagnosis using voice characteristics and machine learning techniques. The human voice is a valuable source of data that provides insight into a person's underlying neurological and physiological state (Neera Batra & Sonali Goyal, 2021) (McFarthing K & Rafaloff G,2022). Vocal alterations in Parkinson's disease are hypothesised to be related to the malfunctioning of the basal ganglia and other brain areas responsible for speech and motor control. Numerous investigations have documented unique vocal impairments in PD patients, such as monotone pitch, slowed speech tempo, and decreased vocal loudness. Even in the early stages of the disease, when motor symptoms are mild or nonexistent, certain voice characteristics may appear. As a result, voice data analysis offers a rare chance to identify PD in its early stages, enabling prompt treatment interventions and maybe delaying the course of the disease (Siddharth Arora et al., 2022). Various medical applications, such as disease detection, image analysis, and drug discovery, have seen impressive success using machine learning. Large datasets of voice recordings from both PD patients and healthy individuals can be used to train machine learning algorithms for the purpose of diagnosing PD. Then, these algorithms can learn to recognise patterns and traits in the voice data that are unique to PD.
This study's organisation is as follows:
The literature on voice-based PD diagnosis is thoroughly reviewed in Section 2 along with the function of machine learning in this situation.
The process used to collect data, extract features, and construct machine learning models is described in Section 3.
The implications of our findings, results of voice-based PD diagnosis using machine learning are covered in Section 4.
Section 5 summarises the relevance of our research and its possible impact regarding early detection and management of Parkinson's disease to finish the research report.