Research On Classification Algorithm Of Extreme Learning Machine Based On Intelligent Optimization

Classification algorithm is one of the important research contents in the field of artificial intelligence and is widely used in various fields of science and engineering.Extreme Learning Machine(ELM)is a single hidden layer Feedforward Neural Network method.During model building,parameters between input layer and hidden layer are randomly initialized,and weights between hidden layer and output layer are obtained through analytical calculation.Compared with traditional neural network model,Extreme Learning Machine model has faster training speed and better generalization ability.Therefore,extreme learning machine as a classifier model,has been widely popular.Based on the research of extreme learning machine,this paper mainly carries out the following work:First,the idea of weighting is introduced into the Two-hidden-layer Extreme Learning Machine,and the Weighted Two-hidden-layer Extreme Learning Machine model is proposed.Two-hidden-layer Extreme Learning Machine not only retains the characteristics of traditional Extreme Learning Machine,but also guarantees the fast training speed of the algorithm.By introducing the idea of weighting,the classification accuracy of the model is further improved.Secondly,Particle Swarm Optimization Algorithm,Moth Fighting Fire Algorithm and Gray Wolf Optimization Algorithm are introduced,and the Improved Gray Wolf Optimization Algorithm is proposed.In the Improved Gray Wolf Optimization Algorithm,the updating formula of convergence factor and position is adjusted to improve the ability of searching optimal parameters.Finally,the weight and bias of input layer to hidden layer in the Two-hidden-layer Extreme Learning Machine model and the Weighted Two-hidden-layer Extreme Learning Machine model are optimized by Particle Swarm Optimization Algorithm,Moth Fighting Fire Optimization Algorithm,Gray Wolf Optimization Algorithm and the Improved Gray Wolf Optimization Algorithm respectively.The influence of the intelligent optimization algorithm on the model classification results was explored by the model classification accuracy and other evaluation indexes.Experimental results show that introducing intelligent optimization algorithm into the model can improve the classification effect of the model under the condition of ensuring the robustness of the model.Meanwhile,the Improved Gray Wolf Optimization Algorithm optimizes the classification effect of the Weighted Two-hidden-layer Extreme Learning Machine model best.Compared with the classification results of Extreme Learning Machine,Regularized Extreme Learning Machine and Two-hidden-layer Extreme Learning Machine,the proposed Weighted Two-hidden-layer Extreme Learning Machine model has the characteristics of good robustness,high accuracy and fast timeliness.The Improved Gray Wolf Optimization Algorithm proposed to optimize the Weighted Two-hidden-layer Extreme Learning Machine model has the best classification effect.Experiments show that the intelligent optimization algorithm can improve the classification performance of the model without changing the structure of the original model.