Use of Artificial Intelligence in Medical Classification for Hemiplegic Patients

Abstract

This study explores a machine learning-based neural network system using MATLAB to classify hemiplegia, a condition causing paralysis on one side of the body. Using data from the specialized treatment center “El laboratorio de marcha en el Hospital Ortopédico Infantil” in Caracas, Venezuela, the study developed an algorithm to categorize patients into four established hemiplegia types. Techniques such as Principal Component Analysis (PCA) and Self-Organizing Maps (SOMs) were used for dimensionality reduction and data clustering, while a Convolutional Neural Network (CNN) refined the classification. The algorithm identified distinct subgroups within the categories, indicating a more complex data structure. Despite promising results in aiding clinical diagnosis, time constraints limited the exploration of these subcategories. This research demonstrates the potential of AI to enhance medical diagnostics, especially in resource-limited settings.