Statistical Distribution of Blood Glucose Levels in Diabetic Patients Diagnosis Using ML-Based PCA Methods

Diabetes mellitus is a chronic illness that is manageable and preventable. It can seriously harm the body and bring a great deal of issues. Therefore, it's critical to receive a diabetes diagnosis as soon as possible and modify your lifestyle to prevent long-term problems from the disease. Blood glucose levels measured after fasting is a crucial first step in an early diagnosis and course of treatment. To predict them, you can utilize information from your medical history. Nonetheless, there is a great deal of dimension, noise, coupling, and nonlinearity in the data found in medical records. The LSSVM, PCA-LSSVM, and LSSVM-KPCA models are the three available. The three models—PCA-LSSVM, LSSVM, and LSSVM-KPCA—are contrasted. KPCA-LSSVM outperforms LSSVM and PCA-LSSVM in terms of accuracy, according to the results. It is the ROC curve's integral. The region is also near one, demonstrating that KPCA-LSSVM is a novel approach to mining medical data and can be used to forecast fasting blood glucose levels.