Automated prognostic tool for cervical cancer patient database

We propose an application, which is developed to assist doctors who treat cervical cancer. The application has a number of features that the previous diagnostic system lacks; particularly, the ability to classify the current stage of cancer from an input pattern. The other features are first the capability of searching for patients whose stages of cancer are the same as the stage of cancer classified from the above input pattern. Second, the ability to search for patients whose considered parameters are similar to the parameters that make up the input pattern. The last feature is to keep the functionality of the previously used system; that is, the capability to retrieve, and display information of patients from a specified patient number. This paper explains the architecture, design, and implementation of the application. The approach employed to classify the stages of cancer is that of a multilayer feed-forward neural networks. Experiments of investigating the architecture of the neural networks to perform the classification task are also given.