Development of an Embedded Device for Stroke Prediction via Artificial Intelligence-Based Algorithms

The most significant cause of disability among the aged and the old is stroke, which causes several social or financial challenges. A stroke may result in death if it is not addressed properly. Patients with such a stroke are often identified with odd bio-signals (i.e., EMG). Therefore, if patients are correctly examined in real-time while being watched and quantified in their physiological signals, they may get the right therapy immediately. However, many stroke screening and diagnostic systems involve pricey and challenging to use in live imaging technologies, i.e., CT or MRI. With the use of real-time Artificial Intelligent (AI) signals, the proposed research created a stroke prediction system that can identify strokes. The system uses Machine Learning (ML) algorithms (XgBoost, Random Forest, Decision Tree, and Voting Classifier). Real-time EMG (Electromyography) bio-signals from the arm and forearm were recorded, after which critical characteristics were released, and prediction models were based on routine activities. It is observed that using the proposed collected data set for four features, the classification accuracy of Random Forest is almost 95%, and it performs better than other classification algorithms such as XgBoost 91%, Decision Tree and Voting Classifier both have (85%).