Smoothed Hinge Loss and $\ell^{1}$ Support Vector Machines
arXiv:1808.071002 citationsh-index: 24
AI Analysis
For practitioners needing scalable ℓ1 SVM solvers, this algorithm offers a practical improvement in computational efficiency.
The paper presents a new algorithm for soft-margin SVM with ℓ1 penalty that requires a modest number of data passes, addressing scalability for large datasets.
A new algorithm is presented for solving the soft-margin Support Vector Machine (SVM) optimization problem with an $\ell^{1}$ penalty. This algorithm is designed to require a modest number of passes over the data, which is an important measure of its cost for very large data sets. The algorithm uses smoothing for the hinge-loss function, and an active set approach for the $\ell^{1}$ penalty.