Diagnosis of Diabetes Using an Intelligent Approach Based on Bi-Level Dimensionality Reduction and Classification Algorithms

Objective: Diabetes is one of the most common metabolic diseases. Earlier diagnosis of diabetes and treatment of hyperglycemia and related metabolic abnormalities is of vital importance. Diagnosis of diabetes via proper interpretation of the diabetes data is an important classification problem. Classification systems help the clinicians to predict the risk factors that cause the diabetes or predict people who are at risk of developing diabetes. Materials and Methods: In this study, Pima Indian diabetes dataset taken from the UCI machine learning repository was used which contains ۷۸۶ samples of normal and diabetes with ۸characteristics.Selection of efficient features of this dataset was analyzed using correlation criterion, information gain and CfsSubsetEval. Then diagnosis of diabetes diseases on Pima dataset was considered using proposed by-level dimensionality reduction method and classification algorithms. Classification algorithms used in this study are KNN, quadratic, Naïve Bayes, nearest mean classifier, non-parametric Gaussian and Mahalonobis kernel and linear discriminant. Results: In all feature selection methods, plasma glucose concentration a ۲-hours in an oral glucose tolerance test, body mass index and age have been selected as the top-ranked features in intelligent diagnosis of diabetes. Proposed method has achieved the accuracy of ۸۲.۰۹ using KNN and quadratic methods and bi-level dimensionality reduction on Pima dataset. The best performance has been achieved by performing PCA algorithm on the features, namely, number of pregnancy, plasma glucose concentration a ۲ hours in an oral glucose tolerance test, body mass index, diabetes pedigree function and Age. Conclusion: The results of this study showed that bi-level dimensionality reduction and classification algorithm scan be very helpful in assisting the physicians to diagnosis diabetes.

کلیدواژه ها:

Diabetes ، Data mining ، Classification ، Dimensionality reduction ، Feature selection ، Feature extraction.