Hybrid Algorithm for Detecting Diabetes
Journal Title: International Research Journal of Applied and Basic Sciences - Year 2014, Vol 8, Issue 12
Abstract
Today's, we can use of new technology tools for restoring and saving data in huge size and so we need a new science for searching in these huge data source and finding useful and necessary results in them. Data mining is a science that is searched automatically in huge data for finding models and association rule in them where other statistical analysis can’t do that. The medical science is one of sciences that need to use of these tools for analyzing their huge data and creating predictive model with new computation ways. The purpose of this research is review of rules and application area of predictive data mining in medical sciences and presenting a frame work for creating, evaluation and exploitation of data mining models in this. In this paper a new method is presented for the diagnosis of diabetes. . Today's foremost practitioners relying on their experience and knowledge, they realize the disease is complex and time consuming experiments. However, human error is inevitable. In this article we want to use the combination of artificial intelligent techniques such as feature selection, decision tree to select the best possible conditions ,KNN ,Neural Network , and education to estimate adaptability learn Machine, in order to identify the diagnosis of uses. This method is compared with conventional approaches one hand and on the other hand, synthetic methods in the references is useful in itself. The present study compared to older methods and meta-heuristic model involving a combination of KNN and neural networks and decision trees and feature selection in the diagnosis of diabetes is based on the laws and properties of the compound.
Authors and Affiliations
Amir Amiri| Department of Computer Engineering, Malayer branch, Islamic Azad University, Malayer, Iran; Email: amiryamir57@gmail.com, Vahid Rafe| Department of computer engineering, Faculty of Engineering, Arak, University, Arak 38156-8-8349, Iran
Keywords
Related Articles
- EP ID EP6880
- DOI -
- Views 368
- Downloads 13
Aryaman god in India & Iran
The god of Airyaman is one of the ancient Hindo-Iranian gods.so first of all I study that in ancient Indian texts(Sanskerit) to know their vision and explanation of this god . He is often mentioned with Mithra and Vruna...
Comparisonof peel volatile componentsof citron and pummelo (Citrus sp)
The peel volatile components of citron and pummelo were investigated in this study.Peel flavor components were extracted by using cold-press and eluted by using n-hexane. Then all analyzed by GC-FID and GC-MS.Fifty-twoan...
Forecasting the Volatility of the Tehran Stock Market Index Using ARCH Models Family
In this study, the function of predicting conditional fluctuation models (5 models) for predicting the total index of Tehran Stock Exchange has been evaluated by using the root-mean-square errors criteria, the mean absol...
SedimentaryStudy of QarranquRiver Using Hec-Ras
Hec-Ras is an applied software package for simulation of river network in steady and unsteady flow regime. The sediment transport simulation is also added in 4.1 version. The goal of this paper is to investigate applicat...
Algorithm Acceleration and Data Storage Volume Reduction in Reliability Modeling within Distribution Networks
Electricity distribution network is among systems reliability of which is of great importance. They are highly complex due to their large size and hence their modeling takes a long time. This paper offers a simple algori...