Generally, the linear kernel and the Gaussian kernel are used, that is, the Linear kernel and the RBF kernel
need to be noted that the data needs to be normalized. Many users forget this small detail
and in general, the RBF effect will not be worse than the Linear
but the time. RBF will cost more, and other students have also explained that the
following is Wu Enda's opinion:
1. If the number of Features is large, similar to the number of samples, then choose LR or Linear Kernel's SVM
2. If the number of Features is relatively small , the number of samples is average, neither too large nor too small, choose SVM+Gaussian Kernel
3. If the number of features is relatively small and the number of samples is large, you need to manually add some features to become the first case
need to be noted that the data needs to be normalized. Many users forget this small detail
and in general, the RBF effect will not be worse than the Linear
but the time. RBF will cost more, and other students have also explained that the
following is Wu Enda's opinion:
1. If the number of Features is large, similar to the number of samples, then choose LR or Linear Kernel's SVM
2. If the number of Features is relatively small , the number of samples is average, neither too large nor too small, choose SVM+Gaussian Kernel
3. If the number of features is relatively small and the number of samples is large, you need to manually add some features to become the first case