Browsing by Author "Çıkrıkçılı, Onur"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Predicting alzheimer’s disease using adaptive neuro fuzzy inference system
    (Bahçeşehir Üniversitesi Fen Bilimleri Enstitüsü, 2013-08) Çıkrıkçılı, Onur; Karahoca, Adem
    Alzheimer's disease (AD) one of the major health problem all around the world and unmitigated cure has not been found yet. A correct diagnosis of AD can be affirmed by histopathologic tests. In addition, mental tests and daily activities can lead diagnose of patients' mental condition. The goal of this study is to develop a data mining solution using neuropsychological test results that makes diagnosis of AD and its stages as accurate as possible and assist to medical doctors' final decision. In this study, Sugeno-Type adaptive-network-based fuzzy inference system (ANFIS), multilayer perceptron (MLP), Iterative Dichotomiser 3 (ID3) and One Rule (OneR) algorithms were assessed whether to could predicting AD. The data set is collected from 264 patients who complained about their health problems and applied to Istanbul University's Department of Neurology. All of the subjects’ ages are 65 or over. The blind data records has 11 attributes that covers basic demographic information and neuropsychological test results. Using “Information Gain” filter, ineffective attributes are eliminated. According to the results, ANFIS classified the instances with the highest correctness rate which is %96 and MLP classified an accuracy of 87%, ID3's is 76% and OneR's is 76%. In addition ANFIS has a high performance based on the methods that sensitivity, specificity and root mean square error.