Advancing Parkinson's Detection Through Intelligent Algorithms

• Published in: 2024 10th International Conference on Advanced Computing and Communication Systems (ICACCS) Vol. 1; pp. 1259 - 1264
• Main Authors: Rasheedha, A., Pratheeba HP, Lynsha Helena, Vinnetia, S. Caroline, Devi, S. Sharmila, Reethika, A., Gunanandhini, S.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.03.2024
• Subjects: Accuracy; Evolutionary computation; Genetic algorithms; handwritten notes; K-Nearest Neighbour; Machine learning algorithms; machine learning techniques; Medical treatment; Parkinson's disease; Parkinson's disease detection; Systematics; Training; Transfer learning
• ISBN: 9798350384352
• ISSN: 2469-5556
• DOI: 10.1109/ICACCS60874.2024.10716973

Parkinson's disease (PD) is a slowly developing nervous disorder that presents with signs like those of other illnesses. Appropriate diagnosis and therapy depend on early detection. One common method for identifying and evaluating Parkinson's disease is handwritten notes. Early Parkinson's disease (PD) detection has been studied using a number of machine learning techniques, although the most have poor performance accuracy. An effective deep learning model is suggested to help diagnose Parkinson's disease early. The model uses evolutionary algorithms, such as the K-Nearest Neighbour approach, to pick the most optimal features for high-performance accuracies. The model obtains an zone under curve of 0.92, detection accuracy of over 96%, and precision of 99%. To illustrate the effectiveness of the model, its performance is contrasted with cutting-edge machine learning and deep learning-based PD detection methods.