Computational intelligence-based diagnosis tool for the detection of prediabetes and type 2 diabetes in India.

摘要: BACKGROUND: The incidence of diabetes is increasing rapidly across the globe. India has highest proportion diabetic patients, earning it doubtful distinction 'diabetes capital world'. Early detection could help to prevent or postpone its onset by taking appropriate preventive measures, including initiation lifestyle changes. To date, early identification prediabetes type 2 proven problematic, such that there an urgent requirement for tools enabling easy, quick, and accurate diagnosis. AIM: develop precise tool diagnosing based on machine learning algorithms. METHODS: dataset used in this study was health profiles non-diabetic patients from hospitals India. A novel algorithm, termed "mixture expert", determination a patient's state. Out total 1415 subjects, 1104 were train mixture expert system. remaining 311 data sets reserved validation algorithm. Mixture implemented matlab development model. model with minimum mean square error selected results. RESULTS: Different combinations numbers hidden nodes expectation maximization (EM) iterations optimize accuracy overall best 99.36% achieved iteration 150 20 nodes. Sensitivity, specificity, classification calculated as 99.5%, 99.07%, 99.36%, respectively. Furthermore, graphical user interface developed java script can readily enter variables easily use algorithm tool. CONCLUSIONS: This describes highly prediction identifying prediabetic, diabetic, individuals high accuracy. be large scale screening hopsitals prevention programs.