A support vector machine method based on chaotic gray wolf optimization

Abstract

The present invention proposes a support vector machine method based on chaotic gray wolf optimization, which specifically combines the chaoticization of the gray wolf algorithm with the support vector machine, and optimizes the two functions of the support vector machine through the chaotic gray wolf algorithm with excellent global generation searching ability. The key parameters are the penalty coefficient C and the kernel width γ to obtain the optimal kernel extreme learning machine parameter value, so that this application can obtain more accurate intelligent decision-making effects, and effectively assist decision-making agencies to make scientific and reasonable decisions, which has important applications value.

Description

technical field [0001] The invention relates to a support vector machine method based on chaotic gray wolf optimization, which belongs to the field of computer science. Background technique [0002] Grid search and gradient descent are currently the two most commonly used parameter optimization methods for support vector machines (SVM). Grid search is an exhaustive search method. It generally divides the specified parameter space by setting reasonable interval upper and lower limits and interval steps, and then trains and predicts the parameter combinations represented by each grid node. These A group of parameters with the highest values ​​in the prediction results are used as the best parameters of the final SVM model. Although this exhaustive search method can guarantee the optimal parameter combination in a given parameter space to a certain extent, as the parameter space increases, its search efficiency will be greatly reduced, especially when setting reasonable interv...

AI Technical Summary

Problems solved by technology

In practical applications, if their values ​​are too large or too small, the generalization performance of SVM will deteriorate.

Images