An Investigation of Blood Sugar Level and Predicting Scrupulous Insistence for Diabetic Recovery Using Data Mining Techniques

Abstract

Diabetes epidemic is one of non-communicable diseases that happens when the pancreas does not produce enough insulin (type 1) or when the body cannot effectively use the insulin (type 2) that needed to regulate glucose. To make controlling and prevention healthcare centers are prepared and implemented several programs. However, it is beyond human ability to analyze huge amounts of data.  To provide an investigation of blood sugar level and predicting scrupulous insistence for diabetic recovery using data mining techniques and MATLAB.  Knowledge discovery in database consisting of nine steps was adopted to extract and discover significant patterns from a dataset of diabetes examination reports. The data was collected from Arba Minch general hospital between the years 2002-2009e.c. and the basic data preprocessing was applied before the data have been used. Clustering using self-organization map and built time series prediction model using nonlinear autoregressive with external inputs was completed. All the correlation plots are indicates that the model is performing correctly in predicting.  And also the training data R= 0.98938 and testing data R=0.99792 this indicates the testing data fit is as good as the training data. In this study illustrated that neural network can be used efficiently to cluster and predict blood sugar cases. Blood glucose test is the primary method for diagnosing diabetes and also can be used as an assistant tool by diabetes specialist to help them to make more consistent diagnosis of diabetes disease.