Abstract:The study presents the concept of a computationally efficient machine learning (ML) model for diagnosing and monitoring Parkinson's disease (PD) in an Internet of Things (IoT) environment using rest-state EEG signals (rs-EEG). We computed different types of entropy from EEG signals and found that Fuzzy Entropy performed the best in diagnosing and monitoring PD using rs-EEG. We also investigated different combinations of signal frequency ranges and EEG channels to accurately diagnose PD. Finally, with a fewer number of features (11 features), we achieved a maximum classification accuracy (ARKF) of ~99.9%. The most prominent frequency range of EEG signals has been identified, and we have found that high classification accuracy depends on low-frequency signal components (0-4 Hz). Moreover, the most informative signals were mainly received from the right hemisphere of the head (F8, P8, T8, FC6). Furthermore, we assessed the accuracy of the diagnosis of PD using three different lengths of EEG data (150-1000 samples). Because the computational complexity is reduced by reducing the input data. As a result, we have achieved a maximum mean accuracy of 99.9% for a sample length (LEEG) of 1000 (~7.8 seconds), 98.2% with a LEEG of 800 (~6.2 seconds), and 79.3% for LEEG = 150 (~1.2 seconds). By reducing the number of features and segment lengths, the computational cost of classification can be reduced. Lower-performance smart ML sensors can be used in IoT environments for enhances human resilience to PD.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to develop an entropy - based machine - learning model for diagnosing and monitoring Parkinson's disease (PD) in the intelligent Internet of Things (IoT) environment. Specifically, researchers use resting - state electroencephalogram (rs - EEG) signals and calculate different types of information entropy to improve the accuracy of diagnosing and monitoring Parkinson's disease. The paper pays special attention to the performance of fuzzy entropy (Fuzzy Entropy) and explores the influence of different signal frequency ranges and EEG channel combinations on the diagnosis effect. In addition, the research also aims to reduce the number of features and the length of signal segments to reduce the computational cost of classification, so that low - performance intelligent sensors can also be used in the IoT environment, thereby enhancing human resistance to Parkinson's disease. ### Main contributions: 1. **Selection of entropy methods**: The research explored the most effective entropy methods to calculate EEG entropy features, among which fuzzy entropy showed the best performance in signal separation between Parkinson's disease and the normal control group. 2. **Frequency range and channel selection**: By identifying the most significant EEG frequency ranges and channel combinations, the research was able to reduce computational complexity while maintaining high diagnostic accuracy. 3. **High accuracy with a low number of features**: A method based on entropy values was proposed to monitor the patient's condition, achieving 99.9% classification accuracy using only 11 features. 4. **Concept of the IoT environment**: A conceptual IoT system was proposed, which can use a machine - learning model to monitor the state of Parkinson's disease patients. By reducing the number of features, the classification cost is reduced, enabling low - performance sensors to be deployed in the health - monitoring system. ### Method overview: - **Dataset**: The research used EEG data from 20 Parkinson's disease patients and 20 age - matched normal control groups. The data was collected by the Emotiv EPOC wireless headset device. - **Signal pre - processing**: The original EEG signals were processed by a fifth - order Butterworth filter to remove low - frequency and high - frequency noise and decomposed into different frequency sub - bands by discrete wavelet transform (DWT). - **Feature generation**: Multiple entropy features were calculated from the pre - processed EEG signals, including singular value decomposition entropy (SVDEn), permutation entropy (PermEn), sample entropy (SampEn), cosine similarity entropy (CoSiEn), fuzzy entropy (FuzzyEn), phase entropy (PhaseEn) and attention entropy (AttnEn). - **Classification evaluation**: Support vector machines (SVM) were used for classification, and hyper - parameters were optimized by repeated K - fold cross - validation (RKF), and finally the classification accuracy was evaluated. ### Experimental results: - **Comparison of entropy methods**: Fuzzy entropy performed the best among all entropy methods, reaching 99.9% classification accuracy. - **Selection of signal types**: Low - frequency signals (especially those in the 0 - 4 Hz range) are particularly important for the diagnosis of Parkinson's disease. - **Channel selection**: EEG channels in the right hemisphere (such as F8, P8, T8, FC6) provided the most informative signals. - **Single - feature classification**: By analyzing a single feature, the most effective frequency range and channel combination were determined, further improving the efficiency and accuracy of classification. In conclusion, this paper proposes an efficient and accurate method for diagnosing and monitoring Parkinson's disease, which is suitable for the IoT environment with limited resources.

Entropy-based machine learning model for diagnosis and monitoring of Parkinson's Disease in smart IoT environment

Parkinsons Disease Detection via Resting-State Electroencephalography Using Signal Processing and Machine Learning Techniques

Machine Learning-Based Detection of Parkinson's Disease From Resting-State EEG: A Multi-Center Study

Parkinson’s Disease Detection from Resting-State EEG Signals Using Common Spatial Pattern, Entropy, and Machine Learning Techniques

Fractal dimensions and machine learning for detection of Parkinson’s disease in resting-state electroencephalography

Parkinson’s Disease Classification Using Machine Learning Approaches and Resting-State EEG

A Hybrid Deep Spatio-Temporal Attention-Based Model for Parkinson's Disease Diagnosis Using Resting State EEG Signals

A novel automated Parkinson's disease identification approach using deep learning and EEG

Diagnosis of Parkinson's Disease Using EEG Signals and Machine Learning Techniques: A Comprehensive Study

A Hybrid Deep Spatiotemporal Attention‐Based Model for Parkinson's Disease Diagnosis Using Resting State EEG Signals

ESDC-LSH: Ensemble Support-Vector Deep Convolutional Based Levy Selfish Herd Optimization for Prediction and Classification of Parkinson's Disease

An efficient Parkinson's disease detection framework: Leveraging time-frequency representation and AlexNet convolutional neural network

Internet of Things Technologies and Machine Learning Methods for Parkinson's Disease Diagnosis, Monitoring and Management: A Systematic Review

An experimental study for early diagnosing Parkinson's disease using machine learning

Classification of Parkinson's disease EEG signals using 2D-MDAGTS model and multi-scale fuzzy entropy

Computational EEG in Personalized Medicine: A study in Parkinson's Disease

Dynamical system based compact deep hybrid network for classification of Parkinson disease related EEG signals

Machine Learning Approaches for Detecting Parkinson’s Disease from EEG Analysis: A Systematic Review

A novel approach for Parkinson’s disease detection using Vold-Kalman order filtering and machine learning algorithms

Statistical feature-based EEG signals classification using ANN and SVM classifiers for Parkinson’s disease detection

Exploring the application and challenges of fNIRS technology in early detection of Parkinson's disease