Todo List

Member print_logo ()

TODO: (by cla) Decide on outpuformat, logging and similar.

Class quickrank::data::QueryResults

TODO: it seems we need also a class withouth features

Member quickrank::io::model_node_to_conditional_operators (pugi::xml_node &nodes, std::stringstream &os)

TODO: this should be changed with item mapping

Class quickrank::io::Svml

TODO: handle feature filtering

Member quickrank::learning::CustomLTR::FIXED_SCORE

TODO: add load_model();

Member quickrank::learning::forests::LambdaMart::fit_regressor_on_gradient (std::shared_ptr< data::VerticalDataset > training_dataset)

TODO: memory management of regression tree is wrong!!!

Member quickrank::learning::forests::Mart::fit_regressor_on_gradient (std::shared_ptr< data::VerticalDataset > training_dataset)

TODO: memory management of regression tree is wrong!!!

Class quickrank::metric::ir::Map

TODO: test correctness

Member quickrank::metric::ir::Metric::jacobian (std::shared_ptr< data::RankedResults > ranked) const

TODO: provide def implementation

Member quickrank::metric::ir::Ndcg::evaluate_result_list (const quickrank::data::QueryResults *rl, const Score *scores) const

TODO: for only zero result slist Yahoo! LTR returns 0.5 instead of 0.0.

Member quickrank::metric::ir::Tndcg::evaluate_result_list (const quickrank::data::QueryResults *rl, const Score *scores) const

TODO: for only zero result slist Yahoo! LTR returns 0.5 instead of 0.0.

Member quickrank::metric::ir::Tndcg::jacobian (std::shared_ptr< data::RankedResults > ranked) const

TODO: it makes sense to pre-compute weights also in ndcg

TODO: jacobian->at is expensive, we should do this in the results list order and not in the re-sorted list

Member RegressionTree::~RegressionTree ()

TODO: memory management of regression tree is wrong!!!

Member RTNode::parse_xml (const pugi::xml_node &split_xml)

TODO: this should be changed with item mapping