Implementation of a Generalized Artificial Intelligence Algorithm Applied to Stroke Prediction in Diabetic Patients.

Thanks to technological advancements, AI and deep learning-based systems have generated significant interest in addressing stroke and diabetes-related issues. In this article, we have, on the one hand, implemented a generalized classification algorithm that combines a set of AI algorithms used for classification, displaying the best model metrics and saving the optimal model. On the other hand, we optimized the computation time of the proposed algorithm. To achieve this, we used the Mojo programming language, which eliminates interpreter overhead and leverages parallelism and hardware optimizations to accelerate computations for all classification algorithms, enabling much faster execution. For real-time classification of stroke risk in both diabetic and non-diabetic individuals, as well as to validate our algorithm, we worked with two datasets: Healthcare and Diabetes.