Harnessing AI Mining Methodologies to Anticipate the Probability of Gestational Diabetes Mellitus Occurrence

The effective diagnosis of medical conditions can be achieved through the application of knowledge discovery techniques in medical databases. Utilizing AI-based mining proves to be a valuable method for extracting insights and uncovering hidden patterns within stored information. This research focuses on utilizing healthcare data from gestational women during the first trimester to predict the risk of gestational diabetes, marking a novel approach in the field. Data preprocessing involves employing an equal width binning interval technique to discretize continuous attributes, with input from medical experts to determine the desired interval width. The discretized values are then used as inputs for the model generation process, which occurs in two phases. Phase 1 involves converting numerical attributes into categorical values through discretization techniques, while Phase 2 applies and compares KNN and CNN algorithms to generate association rules. These rules reveal connections among measured attributes and provide insights into the risk level of gestational diabetes.