Platt: Fast Training of Support Vector Machines using Sequential Minimal Optimization. Note: for improved speed normalization should be turned off when operating on SparseInstances.įor more information on the SMO algorithm, see In the multi-class case, the predicted probabilities are coupled using Hastie and Tibshirani's pairwise coupling method. To obtain proper probability estimates, use the option that fits calibration models to the outputs of the support vector machine. Multi-class problems are solved using pairwise classification (aka 1-vs-1). (In that case the coefficients in the output are based on the normalized data, not the original data - this is important for interpreting the classifier.) It also normalizes all attributes by default. This implementation globally replaces all missing values and transforms nominal attributes into binary ones. Implements John Platt's sequential minimal optimization algorithm for training a support vector classifier.
0 Comments
Leave a Reply. |