Abstract:

Aim: To investigate the ability of neural networks to detect and classify the complete improvement of headache in elderly patients during the follow- up period. Materials and Methods: The multilayer perceptron (MLP), which is the most common neural network, was used to predict prognosis of headache in elderly patients. The data set was divided into training and test sets, and back-propagation algorithm was used as the learning algorithm. The accuracies of the models to predict completely improved patients at the end of 20, 40, and 60 months of follow-up were evaluated by means of the areas under the receiver operating characteristic (ROC) curves. Results: The classification results showed the neural network models had good performance in both training and test phases. In addition, the areas under the ROC curve for each period showed that the accuracies of the models to predict the completely improved patients were in the interval of 0.75-0.90. Conclusions: Neural network model for grouped survival data can be used as a prognostic model. If the prevalence of a disease is low, the sensitivity of the model for detection of the patients with disease will be low.

Keywords: