Parkinson’s disease (PD) is one of the high prevalent neurodegenerative diseases, characterized by motor, non-motor symptoms, and cognitive defects. The disease usually manifests at around the age of 50 years, producing physical, mental, and financial burdens. While current PD therapies rely on symptomatic treatments and clinical observations, there is a driving need for more objective disease-causing genes and biomarkers that can aid in early PD detection. Advances in transcriptomic profiling have led to the exploration of brain tissue-based data of patients, providing insights into the PD pathogenesis. Integration with ML algorithms helps in the early prediction of the disease, leading to appropriate treatments. Our study presents ML classifiers capable of identifying potential disease-causing features, significantly contributing to PD progression. Several classifiers have been used with two feature selection techniques-LASSO and Ridge regression in our prediction model. The performance of each classification model was analyzed using the evaluation metrics, AUC-ROC values, and confusion matrix. The genes produced by LASSO, along with random forest and logistic regression classifiers, were observed as more significant and can be studied further to understand their role in PD pathogenesis.