The challenge to clinician is to diagnosis the complex disease such as Alzheimer, disease, frontotemporal dementia and Parkinson. The diseases are bit complex to diagnose in terms of symptoms as they overlap in many aspects. So, it is necessary to investigate the process of diagnostic with more accuracy with different parameters of the disease. The paper presents the implementation of machine learning algorithms to get more accuracy to identify the disease Parkinson. The data set referred is from Online-based machine learning repository also recognized as UCI. Support vector machine, K-nearest neighbor and linear discriminant analysis algorithms are used to calculate accuracy, recall and confusion matrix. The outcome of this implementation has given the accuracy of 100% for SVM and KNN algorithms and 80% for LDA. The size of the data set is 196 entries, and a number of parameters for each row are 22. The main parameters which are significant for the disease are 5, and implementation has done with 2 different versions. In first version, all the parameters are included, and in the second version, only 5 main parameters included. In both cases, training set is 70% and test set is 30%. SVM shows more accuracy in both versions of implementation. This helps to find the diagnosis and also to judge the significance of parameters with real data set.Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

No abstract