A hybrid approach for early anticipation of brain stroke using ensemble learning

Stroke is a leading cause of mortality and long-term disability worldwide, making early risk assessment critical for effective clinical intervention Being of utmost importance, it becomes the uppermost goal of life to maintain a healthy brain for better health and longevity. It is a sensitive organ that is often vulnerable to many diseases, which at times can have fatal results. One such disease is stroke, which is a result of blockage in brain arteries. Machine learning has gained attention in the medical field, and the medical field has been positively impacted by machine learning. Incorporating an early anticipation system for brain stroke can be useful to prevent the lethal effects of stroke. This study aims to find a robust machine-learning model using the patient characteristics survey dataset for the early prediction of brain stroke. The five machine learning models have been trained by considering a dataset containing sixty-seven characteristics of patients. To find the best classifier, these models have been tested and evaluated based on multiple performance metrics like accuracy, recall, precision, and AUC. The proposed approach based on weighted voting attained an outstanding accuracy of 99.3% along with 99.1% recall, 99.4% precision, and 99.3% AUC score. These results indicate that the proposed approach is highly effective in identifying individuals at risk of stroke and highlights the potential of ensemble learning. The study can be further improved by incorporating multimodal data such as medical imaging to further improve the model.