c# - Can an SVM learn incrementally? -
i using multi-dimensional svm classifier (svm.net, wrapper libsvm) classify set of features.
given svm model, possible incorporate new training data without having recalculate on previous data? guess way of putting be: svm mutable?
actually, it's called incremental learning. question has come before , pretty answered here : a few implementation details support-vector machine (svm).
in brief, it's possible not easy, have change library using or implement training algorithm yourself.
i found 2 possible solutions, svmheavy , lasvm, supports incremental training. haven't used either , don't know them.
Comments
Post a Comment