Training Dendritic Neuron Model with Whale Optimization Algorithm for Classification

Recent research has shown that dendrites are of paramount importance for information processing in neuron. The dendritic neuron model (DNM) was inspired by the effect of dendrites, and constructed via including the nonlinear interactions between excitation and inhibition on dendrites. However, traditional error back-propagation learning algorithm on DNM usually results in the local minima trapping problem, thus limiting its performance. To achieve better performance, we use a nature-inspired meta-heuristic optimization algorithm, i.e., whale optimization algorithm (WOA) which mimics the social behavior of humpback whale to train it. Experimental results are based on several benchmark classification problems. Comparative results indicate that the proposed learning algorithm is more effective for training DNM, thus making DNM more powerful in tackling classification problems.