Gradient Boosted Tree Based Feature Selection and Parkinson’s Disease Classification

Abstract

Parkinson’s Disease (PD) is one of the most common neurodegenerative disorder with increasing prevalence in the world. Many studies have been conducted to detect or diagnose PD. Recently, voice signal based PD detection has been conducted rapidly. However, the performance of the detection still not excellent, hence the need for excellent detection still an open issue in PD detection. In this study, PD detection based on voice signal was conducted using public dataset. Since the dataset consist of many features, the proposed method combines feature selection mechanism and classification based on Gradient Boosted Tree (GBT) algorithm. Feature selection was conducted by using feature weighting from GBT. The experimental results show that the proposed method outperforms prior research with the accuracy about 0.92 with the minimum number of features. This result proves that GBT able to perform feature selection and good classifier. This concludes that the proposed method is promising as PD detection based on voice signal analysis.