Application of Open-Source, Low-Code Machine-Learning Library in Python to Diagnose Parkinson's Disease Using Voice Signal Features

Abstract Parkinson's disease (PD), the second most prevalent neurodegenerative disorder after Alzheimer's disease, affects approximately 10 million individuals worldwide. The disease is characterized by both motor and non-motor symptoms, and clinical aspects are pivotal for diagnosis. Vocal abnormalities can be identified in about 90% of PD patients in the early stages of the condition. Machine Learning (ML), a prominent subfield of Artificial Intelligence (AI), holds significant promise in the medical domain, particularly for early disease detection, enabling effective preventive measures and treatments. In this paper, we considered the unique characteristics of each ML algorithm. Seventeen ML algorithms were applied to a dataset of voice recordings from Healthy Control and PD individuals, sourced from a publicly available repository. We leveraged the PyCaret Python library's ML algorithms and functions, which were introduced in this article, to demonstrate their simplicity and effectiveness in dealing with real-world data. Among these algorithms, Extra Trees Classifier (ETC), Gradient Boosting Classifier (GBC), and K Neighbors Classifier (KNN) exhibited the best performance for the given dataset. Furthermore, to enhance the models' performance, we employed various techniques, including Synthetic Minority Over-sampling Technique (SMOTE) to address class imbalance, feature selection based on correlation, and hyperparameter tuning. Our findings highlight the potential of the PyCaret ML library demonstrated in this article as a valuable tool for applying ML to the classification of Parkinson's disease through voice analysis. The application of ML in this context can greatly support clinical decision-making, leading to more informed and precise interventions.