BudgetedSVM v1.1

Nemanja Djuric, Liang Lan, Slobodan Vucetic, Zhuang Wang — Wed, 12 Feb 2014 20:53:45 -0000

We present BudgetedSVM, an open-source C++ toolbox comprising highly-optimized implementations of recently proposed algorithms for scalable training of Support Vector Machine (SVM) approximators: Adaptive Multi-hyperplane Machines, Low-rank Linearization SVM, and Budgeted Stochastic Gradient Descent. BudgetedSVM trains models with accuracy comparable to LibSVM in time comparable to LibLinear, solving non-linear problems with millions of high-dimensional examples within minutes on a regular computer. We provide command-line and Matlab interfaces to BudgetedSVM, an efficient API for handling large-scale, high-dimensional data sets, as well as detailed documentation to help developers use and further extend the toolbox.

Overview of the main features of the BudgetedSVM package are listed as follows:

We provide efficient implementations of algorithms for highly-scalable non-linear SVM training.
The toolbox can handle large-scale, high-dimensional data sets that cannot be loaded into memory.
The toolbox requires constant memory to train accurate models that solve highly non-linear problems.
We provide command-line and Matlab interfaces to BudgetedSVM.
We provide an efficient API that provides functionalities for handling large-scale, high-dimensional data sets. Using BudgetedSVM API, data sets with millions data points and/or features are easily handled. For more details, please see the documentation included in the download package.

mloss.org BudgetedSVM

BudgetedSVM v1.1