Research Article
Particle Swarm Optimization Based Selective Ensemble of Online Sequential Extreme Learning Machine
Input: training set , learner , trial , threshold | Steps: | (1) for = 1 to | = bootstrap sample from | = | } | (2) generate a population of weight vectors | (3) evolve the population by PSO, where the fitness of the weight vector is defined as | . | (4) = the evolved best weight vector | Output: ensemble : | for regression | for classification |
|