Parameter optimization based two-layer SVM classification model for evaluation of maize breeding

Abstract

In the crop breeding evaluation process, breeders have to observe and record hundreds of thousands of material dates. The traditional breeding evaluation method is effective for selecting the optimal breeding materials by consideration of a lot of characters such as yield, resistance and growth period etc. But the traditional method is difficult to meet the needs of large-scale breeding. Combined with breeders on crop performance evaluation of comprehensive evaluation, the paper proposes parameter optimization based two-layer SVM classification model. The model uses the radial basis function as the kernel function, and uses the method of cross validation to train the sample data several times to obtain the optimal penalty coefficient C and the kernel function parameter g. In the first layer classification model, the breeding trait data is divided into three parts: high yield, stable yield and disease resistance, and the corresponding classification results are obtained. In the second layer model, the first layer classification results are used as the characteristic attribute; being input to the model to get the final category. In order to verify the effect of the model, the paper uses k-neighborhood, decision tree, and random forest and Naive Bayesian classifier as control. The experimental results show that the classification accuracy of the proposed two-layer classification optimization model is 91.523%, is more than other classifiers. So, the parameter optimization based two-layer SVM classification model is suitable for breeding evaluation technology.