DataPoint represents a sparse data point with label.
A linear model stores weights and other information.
Parameter stores the type of solver and user-specified parameters.
A problem stores data points and other necessary information.
Tron is used to solve an optimization problem by a trust region Newton method.
TronFunction defines necessary methods used for different optimization problems in TRON.
TronL2SVM implements TronFunction for L2-regularized L2-loss SVM.
TronLR implements TronFunction for L2-regularized Logistic Regression.