Zhvillimi i Modeleve të Parashikimit të Sëmundjeve në Kujdesin Shëndetësor Duke Përdorur Teknikat e Data Mining

Abstract

This study investigates the use of Data Mining and Machine Learning approaches to support the early detection of cardiovascular diseases, which continue to represent a major concern in healthcare systems worldwide. The research focuses on designing and evaluating multiple predictive models based on the publicly available UCI Heart Disease dataset (Cleveland), with the aim of identifying patterns that can assist in improving diagnostic decision-making. Utilizing a Python-based development environment, advanced data preprocessing procedures were implemented, including missing value imputation and feature standardization. Two primary algorithms were analyzed and compared in this study: Logistic Regression and Random Forest. The experimental results revealed that the Random Forest model achieved a high level of accuracy with an accuracy of 88.52%, outperforming Logistic Regression, which recorded 86.89%. Beyond accuracy, the confusion matrix analysis confirmed that Random Forest provides higher clinical safety by minimizing "False Negatives" (only 1 case). The findings of this study confirm that integrating Data Mining techniques into clinical decision-making systems can significantly improve early diagnosis and the management of patients at cardiac risk.