Threat assessment of support vector machine optimized by Krill Herd algorithm

To put forward the method of threat assessment of support vector machine optimized by Krill Herd algorithm based on the traditional support vector machine optimization method, so as to improve the forecast precision of target threat assessment. The thesis introduces the principles of Krill Herd algorithm and support vector machine and optimize the penalty parameter and kernel function parameter in the support vector machine with Krill Herd algorithm to find the optimal penalty parameter and kernel function parameter; establishes the model of target threat assessment of the support vector machine optimized by Krill Herd algorithm and achieves the target threat assessment algorithm based on this model. Collect 90 sets of original data to form the training set and 30 sets of data to form the test set to carry out simulation experiment on the target threat assessment algorithm. The experimental result shows that the forecast error of the support vector machine optimized by Krill Herd algorithm is 0.002 91 which is less than that of the support vector machine optimized by particle swarm algorithm or firefly algorithm. It can conclusion that the target threat assessment method of the support vector machine optimized by Krill Herd algorithm can effectively complete the target threat assessment. © 2016, Science Press. All right reserved.