Parkinson's disease (PD) is a neurodegenerative condition and a progressive nervous system disorder condition, marked by its characteristic symptoms such as tremors, muscle stiffness, and slow, imprecise body movements. Typically affecting the middle-aged and elderly, the most crucial challenge in the treatment of PD is posed in the form of its early diagnosis. While there is no cure for the disease, early intervention is found to have alleviated the symptoms and to have also improved the overall quality of life, potentially slowing down the disease’s progression. With the rapid advancements in Artificial Intelligence (AI) and Machine Learning (ML), almost every industry seems to be benefiting from its incorporation and the field of biomedical sciences is no different. AI-ML techniques have seemed to be specifically beneficial in classification-based biomedical problems. This research is built upon this very same idea. The application of 14 ML algorithms is studied on a dataset that consists of vocal characteristics of several individuals, both the healthy and the PD patients, to classify based on a patient’s speech whether or not, they have Parkinson’s. To improve the overall performance of the algorithms, SMOTE and PCA pre-processing techniques have also been applied to deal with the dataset’s shortcomings. The study observes that Random Forest, Gaussian Process, Extra Trees, and CatBoost are the 4 best performing algorithms for the task, providing an accuracy of 96.61%. The paper concludes with the deployment of an Explainable AI (XAI) LIME module on the best-performing algorithms to provide transparency and interpretability to the decision-making process of these AI systems.