An improved GA-SVM algorithm

Support vector machine (SVM) can ensure the promotion capability of machine model, so it is widely used in various fields. The selection of SVM´s parameters has a great effect on its performance, if genetic algorithm (GA) is introduced to optimize support vector machine´s parameters, the effect will be better. Traditional GA-SVM algorithm can optimize SVM parameters including penalty factor C and radial basis kernel function parameter σ but other parameters can also affect its performance. In order to improve prediction accuracy, the loss function parameters ε is introduced in this article based on traditional GA-SVM algorithm. Then, copy operator, crossover operator and mutation operator of GA are optimized. Using the traditional GA-SVM algorithm and the improved GA-SVM algorithm to predict concrete compressive test data respectively, the experimental results show that the improved GA-SVM algorithm can significantly improve the accuracy of the data.