Smart Parkinson’s Disease Detection System Using AI-Based Voice Pattern Analysis

Abstract— Parkinson's Disease (PD) is a neurodegenerative disease affecting the human movements, speaking, and motor control functions. The early diagnosis of Parkinson's Disease is of extreme importance to treat and enhance the quality of life of patients suffering from this disease. Currently, the diagnosis of Parkinson's Disease occurs mainly through observations, which detect the disease in the later stages. In the speech signal, hidden patterns are embedded, which can be used for the early detection of Parkinson's Disease. This project proposes a machine learning-based system for the early detection of Parkinson's Disease using speech signals, which are characterized by frequency, intensity, and time variations. In the proposed system, advanced machine learning algorithms, namely, Convolutional Neural Networks (CNN), Support Vector Machines (SVM), and Random Forest, are used for the classification of the speech signal as Parkinson's or normal. In addition, the proposed system uses Gemini AI for improving the performance of the machine learning algorithms in terms of accuracy and adaptability with new data. The proposed system includes speech pre-processing, feature extraction, model training, and evaluation to ensure accurate prediction results. The performance of the proposed model is evaluated using standard metrics such as accuracy and classification report. The experimental results prove the efficiency of the machine learning models and the optimization algorithms to achieve high accuracy for the diagnosis. The objectives of the project are to design an efficient, scalable, and intelligent system for the early detection of Parkinson’s Disease. The proposed system can help the medical professionals to make faster and accurate diagnosis decisions. Keywords— Parkinson’s Disease, Speech Signal Analysis, Machine Learning, Convolutional Neural Networks (CNN), Support Vector Machine (SVM), Random Forest, Early Diagnosis.