Detection of Parkinson's Disease via J48 Algorithm


Çetinkaya S. , Demir A. , Kodal Sevindir H.

10th International Conference on Image Processing, Wavelet and Applications (IWW2019), Kocaeli, Turkey, 18 - 20 October 2019, pp.58

  • Publication Type: Conference Paper / Summary Text
  • City: Kocaeli
  • Country: Turkey
  • Page Numbers: pp.58

Abstract

Machine Learning is used in many areas around the world, including the healthcare industry.It can play an essential role in predicting the presence/absence of neurodegenerative disorder. If this information is well predicted, it may help doctors to treat patients. Parkinson’s Disease is a worldwide health problem, causing movement disorder and gait deficiencies. In this study, walking force signals from the force sensitive receptors were used to separate one of the neurodegenerative Parkinson’s Disease (PD) from the normal individual. These signals were first separated by the Daubechies 4 (db4) wavelet transform to the seventh level. Then peak analysis was performed on the seventh level approximation signal to find out the local maxima of the signal, the peak positions of these local maxima, peak widths, and peaks. Then, two basic statistical characteristics were obtained from each of these four peak features. So, a total of 16 attributes were obtained, 8 for the left foot, and 8 for the right foot. After this stage, training data was generated by using 50% of the data of each subject. In the next step, 87.1% accuracy was obtained for determination of PD using the J48 algorithm.