ModulesΒΆ
Regression and classification:
- RLS - regularized least-squares regression
- LeaveOneOutRLS - RLS with leave-one-out regularization parameter selection
- KFoldRLS - RLS with Kfold cross-validation regularization parameter selection
- LeavePairOutRLS - RLS with leave-pair-out regularization parameter selection
- CGRLS - linear conjugate gradient RLS
Learning to rank:
- GlobalRankRLS - ranking regularized least-squares, ordinal regression
- QueryRankRLS - ranking regularized least-squares, query-structured data
- PPRankRLS - ranking regularized least-squares, pairwise preferences
- LeavePairOutRankRLS - GlobalRankRLS with leave-pair-out regularization parameter selection
- KfoldRankRLS - GlobalRankRLS with Kfold cross-validation regularization parameter selection
- LeaveQueryOutRankRLS - QueryRankRLS with leave-query-out regularization parameter selection
- CGRankRLS - linear conjugate gradient RankRLS
- PCGRankRLS - linear conjugate gradient RankRLS with pairwise preferences
Feature selection:
Paired inputs, Kronecker kernels, zero-shot learning:
Clustering:
Kernels:
Performance measures:
- accuracy - Binary classification accuracy
- auc - Area under ROC curve
- cindex - Concordance index (pairwise ranking accuracy)
- fscore - F1-score
- ova_accuracy - one-vs-all multiclass accuracy
- spearman - Spearman rank correlation
- sqmprank - squared magniture preserving ranking error
- sqerror - mean squared error
Predictors: