Hybrid algorithm based on K-nearest-neighbor algorithm and Adaboost with selection of feature by genetic algorithms for the diagnosis of diabetes

iabetes is known to be one of the chronic and dangerous diseases the people in afflicted with which are increasing in number. On the other hand, early detection of this disease can be effective in treatment and control of disease progression for patients. Using the data mining techniques and machine learning, disease detection systems have been able to help patients and physicians to detect the disease in due time. In this paper, a hybrid algorithm based on the algorithm of K nearest neighbors and Adaboost has been proposed to improve the accuracy of detection of diabetes. In the first stage, the genetics algorithm is used to select the features and the dimensions of feature vector are reduced. Then, the combination of Adaboost algorithm and K nearest neighbor algorithm having been used, the samples are classified. To evaluate the algorithm, it was tested on PID dataset by 10-fold cross validation. The results obtained indicate the high accuracy and speed of the proposed method.