Leveraging Management Information Systems(MIS) and Machine Learning for Predictive Diabetes Risk Assessment

Abstract

Management information systems (MIS) coupled with artificial intelligence (AI) have received much interest in the healthcare sector, and specifically in predictive health management. The study utilizes health-related variables like blood glucose levels, body mass index, age, and pregnancy count to examine the potential application of machine learning models to predict the possibility of diabetes. The study will provide better healthcare decision-making through the early diagnosis and treatment of diabetes. The choice of Random Forest and Logistic Regression machine learning algorithms was based on the fact that they can handle intricate data and can also produce reliable risk estimates. The data that was studied included common health factors that appear in Electronic Health Records (EHR). Two of the examples of data preparation operations that strengthened the models were normalization and missing values. Random Forest had better results in all the key performance metrics, such as ROC-AUC (0.98 vs. 0.82), recall (87.90% vs. 83.20%), accuracy (85.60% vs. 81.50%), and precision (84.20% vs. 80.10%), than Logistic Regression. It has been revealed that the Random Forest model is more successful in determining the diabetes risks, which is essential to early intervention and prevention. Based on these findings, the application of AI-based devices in the healthcare management information system (HMIS) is increasingly becoming useful. These machine learning algorithms can significantly increase early detection of diabetes, decrease misdiagnosis, and offer more specific treatment that will benefit patients and save the healthcare system.

Ähnliche Arbeiten