Evolutionary Extreme Learning Machine Based on Particle Swarm Optimization and Clustering Strategies

Extreme Learning Machine (ELM) is a learning method for single-hidden layer feedforward neural network (SLFN) training. ELM approach increases the learning speed by means of randomly generating input weights and biases for hidden nodes rather than tuning network parameters, making this approach much faster than traditional gradient-based one. In this paper, a hybrid ELM and Particle Swarm Optimization (PSO) approach is presented to optimize the input weights and hidden biases for ELM, which also use the concepts of Clustering Analysis. Two different treatments are presented for the particles that fly out the search space bounds. Experimental results show that the proposed method is able to achieve better performance than ELM for real benchmark datasets.