Study of Dynamic Group Evolution for Health Prediction of Mangrove Ecosystem

摘要

Many researchers have applied neural networks to mangrove ecosystem health research. However, it is challenging to find efficient training algorithms for neural networks, so the prediction results are often difficult to meet the needs. The dynamic group evolution (DGE) algorithm is a recently proposed metaheuristic algorithm, which exhibits a rapid convergence rate and good performance in searching and avoiding local optima. In the present study, we use DGE and back propagation (BP) for training feed-forward neural network and build a DGBPNN algorithm model, which takes into account both global exploration and local exploitation. In this paper, we use the model for man-grove ecosystem health prediction, and the performance of the proposed algorithm is determined by comparing its performance with those of other optimization training algorithms. The experimental results show that our proposed training algorithm exhibits promising performance in training neural networks and is expected to apply to more real-world problems.