Second-Order SMO Improves SVM Online and Active Learning

Iterative learning algorithms that approximate the solution of support vector machines (SVMs) have two potential advantages. First, they allow for online and active learning. Second, for large datasets computing the exact SVM solution may be too time consuming and an efficient approximation can be preferable. The powerful LASVM iteratively approaches the exact SVM solution using sequential minimal optimization (SMO). It allows for efficient online and active learning. Here, this algorithm is considerably improved in speed and accuracy by replacing the working set selection in the SMO steps. A second order working set selection strategy, which greedily aims at maximizing the progress in each single step, is incorporated.

• Type: Journal
• Year: 2008
• Where: NECO
• Authors: Tobias Glasmachers, Christian Igel
• Keywords: Exact Svm Solution, Iterative Learning Algorithms, NECO 2008, Support Vector Machines