Abstract
Backpropagation is a well-known learning algorithm used to train Multilayer Perceptron (MLP) with the iterative process. However, one of the critical shortcomings with the BP learning strategy is that it can sometimes trapped in the local minima with suboptimal weights due to the existence of many local optima in the solution space. To remove the above drawbacks of BP algorithm, researchers interested in naturally swarm intelligence algorithm. Here Guided Artificial Bee Colony (GABC) and Gbest Guided Artificial Bee Colony (GGABC) algorithms are hybrid called Hybrid Guided ABC algorithm. The HGABC algorithm used to train MLP for earthquake time-series data for the prediction task. The performance of proposed Hybrid Guided ABC compared with ordinary BP, GGABC and ABC learning algorithms. The simulation results show that proposed learning algorithm HGGABC has outstanding prediction performance than BP, GGABC and ABC algorithms.