: This paper proposes twin prototype support vector machine (TVM), a constant space and sublinear time support vector machine (SVM) algorithm for online learning. TVM achieves its ...
Support Vector Machines (SVM) are one of the most useful
techniques in classification problems. One clear example
is face recognition. However, SVM cannot be applied
when the fe...
This paper develops algorithms to train support vector machines when training data are distributed across different nodes, and their communication to a centralized processing unit...
Pedro A. Forero, Alfonso Cano, Georgios B. Giannak...
Based on studies and experiments on the loss term of SVMs, we argue that 1-norm measurement is better than 2-norm measurement for outlier resistance. Thus, we modify the previous 2...
We study a class of algorithms that speed up the training process of support vector machines (SVMs) by returning an approximate SVM. We focus on algorithms that reduce the size of...