Onset Diabetes Diagnosis Using Artificial Neural Network

Diabetes Mellitus is a chronic, lifelong metabolism disorder that affects the ability of the body system to use the energy found in food. People living with high blood sugar will experience polyuria (frequent urination), which will make them to become increasingly thirsty (polydipsia) and hungry (polyphagia) .The improper management of this disease can lead to complication such as cardiovascular disease, kidney disease, eye disease, nerve disease, pregnancy complication. The database of Pima Indian diabetes has been considered for the diagnosis of the diabetes mellitus. This database comprise of certain attributes which are very adequate for diabetes mellitus diagnosis. The use of this attributes has enhanced the training and test classification of patients, whether diabetes is present or not. In this research work, multilayer feed-forward was created and trained with back-propagation algorithm which classify patient that are tested positive as binary 1 and patient that are tested negative as binary 0.The use of trained neural network gave recognition rate of 82% on test .This recognition rate was later compared to previous researches on diabetes where other types of algorithms was used such as ADAP algorithm, C4.5 algorithm, nearest neighbor with backward sequential selection of feature, EM algorithm. The success rate obtained from multilayer feed-forward trained with back-propagation algorithms is higher than these other algorithms. Index Terms – Artificial neural network, Back propagation, diagnosis, diabetes. ——————————  ——————————