System and method for boosting support vector machines
AssigneeInternational Business Machines Corporation
InventorsByron Edward Dom, Jianchang Mao, Dmitry Pavlov
Life of this patent
Estimating the term…
Abstract
A system and method for training an SVM in a scalable manner includes boosting the SVM during training. Specifically, individual SVMs in an ensemble of SVMs are trained using small subsets of a training data set, with data that earlier classifiers in the ensemble incorrectly classified being overrepresented in succeeding subsets. In this way, the speed with which the overall SVM is trained is increased and the memory requirements therefor are reduced, even for relatively large training data sets.
Claims17
A computer-implemented method for training a support vector machine (SVM) useful in machine learning, comprising the acts of:
- boosting the SVM.
A computer programmed to undertake logic for providing a support vector machine (SVM), the logic comprising:
- inputting successive subsets of training data into successive SVMs to obtain a final SVM having iteratively derived parameters of an optimal hyperplane.
A computer program product including:
- computer readable code means for receiving at least one support vector machine (SVM);
- and
- computer readable code means for boosting the SVM using iterations of training data subsets.
The computer program product of clam 14, wherein the subset of data is based at least in part on data in a prior subset.
Dependencies are inferred from the claim text.