# An Improved Approach for Prediction of Parkinson's Disease using Machine Learning Techniques

**Authors:** Kamal Nayan Reddy Challa, Venkata Sasank Pagolu, Ganapati Panda, Babita Majhi  
**arXiv:** [1610.08250](https://arxiv.org/abs/1610.08250)  
**Date:** 2016-10-26  
**Summary page:** https://www.summarizepaper.com/en/paper/1610.08250v1/

## Key points

- Parkinson's disease (PD) is a global public health issue affecting millions of individuals worldwide.
- Early prediction of PD is crucial for timely treatment plans.
- Recent research focuses on predicting PD based on non-motor symptoms that precede motor symptoms.
- Identifying these non-motor symptoms early allows for appropriate treatment at the right time.
- This study aims to develop machine learning models to predict PD using non-motor features such as REM sleep Behavior Disorder (RBD) and olfactory loss.
- Important biomarkers are incorporated into the study.
- Various machine learning models, including Multilayer Perceptron, BayesNet, Random Forest, and Boosted Logistic Regression, were used in the study.
- Boosted Logistic Regression achieved the best performance with an accuracy of 97.159% and an area under the ROC curve of 98.9%.
- Machine learning models can effectively be employed for early prediction of PD based on non-motor symptoms.
- Early detection has significant implications for improving patient outcomes through timely intervention and treatment planning.

## Summary

Parkinson's disease (PD) is a significant global public health issue affecting approximately one million individuals in the United States and around five million people worldwide. Early prediction of PD is crucial for initiating timely treatment plans. While most people are familiar with the motor symptoms of PD, recent research has focused on predicting the disease based on non-motor symptoms that precede the motor symptoms. By identifying these non-motor symptoms early on, patients can receive appropriate treatment at the right time. This study aims to develop machine learning models that can accurately predict PD using non-motor features such as Rapid Eye Movement (REM) sleep Behavior Disorder (RBD) and olfactory loss. Additionally, important biomarkers are incorporated into the extended work. The researchers utilized various machine learning models that have not been previously used for this purpose, including Multilayer Perceptron, BayesNet, Random Forest, and Boosted Logistic Regression. The results indicate that Boosted Logistic Regression achieved the best performance with an impressive accuracy of 97.159% and an area under the Receiver Operating Characteristic (ROC) curve of 98.9%. These findings suggest that these machine learning models can be effectively employed for early prediction of PD. Overall, this study highlights the importance of early detection and provides evidence for the potential use of machine learning techniques in predicting PD based on non-motor symptoms. The findings have significant implications for improving patient outcomes through timely intervention and treatment planning.