Comparison of various classification algorithms in the diagnosis of type 2 diabetes in Iran

Abstract

In today's medical world, data on symptoms of patients with various diseases are so widespread, that analysis and consideration of all factors is merely not possible by a person (doctor). Therefore, the need for an intelligent system to consider the various factors and identify a suitable model between the different parameters is evident. Knowledge of data mining, as the foundation of such systems, has played a vital role in the advancement of medical sciences, especially in diagnosis of various diseases. Type 2 diabetes is one of these diseases, which has increased in recent years, which if diagnosed late can lead to serious complications. In this paper, several data mining methods and algorithms have been used and applied to a set of screening data for type 2 diabetes in Tabriz, Iran. The performance of methods such as support vector machine, artificial neural network, decision tree, nearest neighbors, and Bayesian network has been compared in an effort to find the best algorithm for diagnosing this disease. Artificial neural network with an accuracy rate of 97.44 % has the best performance on the chosen dataset. Accuracy rates for support vector machine, decision tree, 5-nearest neighbor, and Bayesian network are 81.19, 95.03, 90.85, and 91.60 %, respectively. The results of the simulations show that the effectiveness of various classification techniques on a dataset depends on the application, as well as the nature and complexity of the dataset used. Moreover, it is not always possible to say that a classification technique will always have the best performance. Therefore, in cases where data mining is used for diagnosis or prediction of diseases, consultation with specialists is inevitable, for selecting the number and type of dataset parameters to obtain the best possible results.

Machine Learning Methods for Diabetes Prediction

“…The model was examined by comparing the random forest (RF) and the normal SVM, and the results show that WVKSVM has better prediction ability to improve the performance of SVM classifier when utilizing the metabolomics datasets. The best method to diagnose T2DM has been investigated using data from Tabriz, Iran [17]. The algorithm such as support vector machine, artificial neural network, decision tree, nearest neighbors, and Bayesian network was chosen to handle diagnose of T2DM.…”

Statistical and Machine Learning Methods for the Classification of Type -2 Diabetes Mellitus

“…The complex data with more parameters may be suitable for one model and unsuitable for other. Feature and parameter selection needs experienced expert's help to increase efficacy [10]Usually Weka softwear , which is available free , is used Wenqian Chen et al Some times a single model may not give as good performance as combination of two or more models ,each model doing separate function one after other ,for example first may be used for data correction followed by another for classification [11]. The professionals and scientists are working together to develop an automatic machine to diagnose a disease especially diabetes and its early prediction in non diabetics .The accuracy of these machine models varies The study on Indian population shows satisfactory results [12] It is observed that on going clinical trials and patients registered with health care organisations all over globe generate huge data which may be unstructured and even fixed data .To utilize this useful information for patient care management, one has to change from older methods to machine learning effective tool.…”

Differential Diagnosis Model of Hypocellular Myelodysplastic Syndrome and Aplastic Anemia Based on the Medical Big Data Platform

“…The results show that the SVM has the highest classification accuracy. Heydari et al [18] compared neural network, SVM, decision tree, and Bayesian methods in the diagnosis of type 2 diabetes and found that the highest accuracy of the neural network model is 97.44%, the decision tree is 95.03%, and the Bayesian network is 91.60%, while the accuracy of SVM is only 81.19%. Lui et al [19] used SVM, Bayesian networks, radial basis neural networks, and multilayer perceptrons to establish a classification model of magnetic resonance features of mild traumatic brain injury.…”