Covariate Shift by Kernel Mean Matching
Published on Jan 19, 20109186 Views
Given sets of observations of training and test data, we consider the problem of re-weighting the training data such that its distribution more closely matches that of the test data. We achieve this g