Optimizing Neural Networks using Dragonfly Algorithm for Medical Prediction

Abstract

As the early diagnoses helps in curing most diseases or in making them more bearable, it increases the need to build a good prediction model. Artificial Neural Network (ANN) is one of the evolutionary computation techniques that can be used as a prediction model for new data records. The ANN training method implemented is backpropagation which can be used in conjunction with an optimization method such as Dragonfly Algorithm (DA) and Artificial Bee Colony (ABC) algorithm. ABC has been used in the optimization of synaptic weights from an ANN. This research presents the use of DA as an optimization algorithm for the weights of each connection between the ANN neurons in a model called (ANN-DA). A disease prediction model that makes use of ANN-DA and takes real data as a case study is presented. We shall also present a model that makes use of ABC, ANN-ABC, in order to evaluate ANN-DA. An experimental comparison of both optimization algorithms with other well-known classifiers has been made using different metrics. The results show that ANN-DA proved its efficiency over ANN-ABC and regular ANN. Furthermore, ANN-DA results were more robust for all datasets.