Frank-Wolfe for Sign-Constrained Support Vector Machines

Domain knowledge is useful to improve the generalization performance of learning machines. Sign constraints are a handy representation combine domain with machine. In this paper, we consider constraining signs weight coefficients in linear support vector machine, and develop an optimization algorithm for minimizing empirical risk under sign constraints. The based on Frank-Wolfe method that also converges sublinearly possesses clear termination criterion. We show each iteration requires O(nd+d2) computational cost. Furthermore, derive explicit expression minimal number ensure ε-accurate solution by analyzing curvature objective function. Finally, empirically demonstrate promising technique when similarities training examples compose feature vector.