Multi-factors correlation to diabetes using Machine Learning: findings from BRFSS

Abstract

The escalating prevalence of diabetes has emerged as a burgeoning national concern, necessitating comprehensive investigation and intervention. Leveraging data extracted from the 2021 Behavioral Risk Factor Surveillance System (BRFSS), this study harnesses the power of data mining techniques to discern salient variables within expansive datasets. Subsequently, a series of machine learning models, comprising Random Forest, Decision Tree, Lasso Regression, and Neural Network, were constructed, and evaluated. The principal objective of this research is to facilitate the identification of the optimal predictive model for diabetes. By scrutinizing and contrasting the performance of these diverse models, this study aspires to contribute valuable insights into the field of diabetes prediction, potentially aiding in the development of more accurate and effective diagnostic tools. Consequently, this research and to make a meaningful stride towards mitigating the adverse impacts of the burgeoning diabetes epidemic and underscores the pivotal role of data-driven analytics and machine learning in the realm of public health research and policy formulation.