Cial Neural Networks in Prediction of Essential Hypertension

This paper investigates the ability of several alternate models of Artificial Neural Network (ANN) to predict the probability of occurrence of Hypertension (HT) in a mixed patient population. To do this a two-layer feed-forward neural network with 13 inputs and 1 output was created with a single hidden layer . Different types of networks structures, such as NEWFF (feed-forward back propagation network), NEWCF (cascade-forward back propagation network) and NEWELM (element back propagation network) were coded and tested. The Levenberg-Marquardt back propagation algorithm was used to train the network. A detailed database, comprising healthy and hypertensive patients from a university hospital was used for training the ANN and prediction. All three network structures tested showed reasonably good accuracy in prediction of disease (s), with NEWEF showing best prediction in 3 out of 4 datasets and NEWCF in one database. Thus the best choice appears to be situation specific.