Early detection of severity of any kind of disease is anessential factor. This helps in treating the patient well ahead. This paperintended to verify the effectiveness of the application of Deep Learning forpredicting the severity of Parkinson’s disease in a patient using his or hervoice characteristics. The dataset used was UCI’s Parkinson’s TelemonitoringDataset, comprising of 16 attributes or biomedical voice measurements withvarious range of values from 42 people with early-stage Parkinson’s disease. Itwas first pre-processed by applying normalisation. Then the segmentation of thenormalised dataset was done to create training dataset and testing dataset.Deep Neural Networks were trained on the training data, and finally theaccuracy of severity prediction was obtained by running the network on thetesting data. We were able to successfully implement deep neural network inpredicting the severity of Parkinson’s disease, achieving an accuracy of 81.
6 %and 62.7 % in the case of motor-UPDRS and total-UPDRS scores respectively. Inorder to analyse the dataset and make an attempt to understand the trend of theseseverity scores, an analysis of the normalised dataset was performed on thebasis of gender and age of patients. The results indicate that accurateprediction of severity of Parkinson’s disease can be done using deep learning.This implies that Deep Learning can be used for severity prediction and medicalanalysis for other diseases of similar types as well.
Although we have used adataset of 5875 instances, the accuracy of our approach can be further improvedby implementing it on a larger dataset, having more number of instances of eachseverity class. Moreover, more number of patient attributes like- gait andhandwriting features- can be added to make the model more reliable. Also, morepowerful computing resources(i.e. GPUs with better processing capabilities) canbe used to improve the time complexity of our approach.