ABSTRACT

Insulin deficiency would be a feature of hyperglycemia Diabetics Type (DT), which was among the most frequent metabolic disorders. The inability of glucagon to function properly creates unmanageable blood sugar levels in the body that could lead to the existing conditions. As a consequence, early intervention of DT was critical to saving many lives. This paper provides a computer learning-based forecasting approach for detecting DT to achieve this goal. The PIMA Indian Women Mellitus database was used to create the classification models, which used the Q-learning technique from the Reinforcement Algorithm (RL) methodology. A model generates an over predicated RL and instructs that learning operative discover to best strategy of three criteria’s (such as weight, insulin level, and patient age) to diagnose patients with DT. A subject's data could be in any of 330 multiple states. The reliability, specificity, memory, F-measure, and AUC values of the suggested RL study were compared to those of government approaches like K Nearest Neighbor (KNN) suggested in terms of effectiveness. Keywords: Hyperglycemia; Diabetics Types; Reinforcement Learning; Decision Tree; K nearest Neighbor
Publication date: 01/11/2021

https://ijbpas.com/pdf/2021/November/MS_IJBPAS_2021_NOV_SPCL1102.pdf

Download PDF
https://doi.org/10.31032/IJBPAS/2021/10.11.1102