C is the penalty coefficient, which is understood as adjusting the preference weight of two indicators (interval size and classification accuracy) in the optimization direction, that is, the tolerance of errors. The higher c, the more intolerable errors, the easier it is to over-fit, the smaller c, the easier it is to under-fit, and the larger or smaller c, the worse the generalization ability.

γ is a parameter when RBF function is selected as kernel. It implicitly determines the distribution of data after mapping to the new feature space. The greater the gamma, the fewer the support vectors, and the smaller the gamma value, the more the support vectors. The number of support vectors affects the speed of training and prediction.

Extended data:

1, support vector machine (SVM) is a kind of generalized linear classifier that classifies data by supervised learning, and its decision boundary is the hyperplane with the largest side distance of learning samples.

2.SVM uses hinge loss function to calculate empirical risk, and adds regularization term to the solution system to optimize structural risk. This is a sparse and robust classifier. Kernel method is a commonly used kernel learning method, and SVM can be used for nonlinear classification.