Prediction of Parkinson Disease in Unsupervised Way using Hybrid Approach through MRI Images

With the help of the Parkinson's Telemonitoring dataset, this paper presents a deep neural network model for predicting the progression of Parkinson's disease. As a result of Parkinson's disease (PD), body movement becomes increasingly difficult. This rating scale is used to determine whether or not someone has PD (UPDRS). To deal with the dataset's multicollinearity issues and reduce the input feature space's dimension, principal component analysis (PCA) is then used. K-clustering models are then used to predict sensitivity, specificity, and processing time for patients with Parkinson's disease using the reduced features fed into the k-clustering model. The proposed model has computed the sensitivity of 99.78%, specificity of 98.16% and the processing time of 6.4 s, which is better than other existing algorithms such as ANN, and mRMR merged with CVANN.