matrix with a sparse vector.

like normal rather than passing it back by reference.

unit test code.

it interprets its epsilon parameter the same way the other OCA based SVM tools
in dlib do.  It also has a more intuitive interpretation than the previous
stopping condition.

ranking_pair.  Also improved wording of comment in svm_rank_trainer_abstract.h

sparse_linear_kernel.

since it might be expensive to call depending on how the user implements it.

user friendly.

serialization.