[mlpack-svn] [MLPACK] #350: transition to arma::Row<size_t> for holding labels
MLPACK Trac
trac at coffeetalk-1.cc.gatech.edu
Mon May 19 09:55:35 EDT 2014
#350: transition to arma::Row<size_t> for holding labels
--------------------------------------------------------------------------------+
Reporter: rcurtin | Owner:
Type: defect | Status: new
Priority: minor | Milestone: mlpack 1.1.0
Component: mlpack | Resolution:
Keywords: arma::Col<size_t>, arma::Row<size_t>, labels, supervised learning | Blocking:
Blocked By: |
--------------------------------------------------------------------------------+
Description changed by rcurtin:
Old description:
> Often in mlpack's supervised learning methods, the type
> `arma::Col<size_t>` is used to hold labels. However, data is column-
> major, which means each dimension is a row. Therefore, it makes sense
> that labels should also be a row, and the type `arma::Row<size_t>` should
> be used instead.
>
> The reason that `arma::Col<size_t>` was originally used was convenience.
New description:
Often in mlpack's supervised learning methods, the type
`arma::Col<size_t>` is used to hold labels. However, data is column-
major, which means each dimension is a row. Therefore, it makes sense
that labels should also be a row, and the type `arma::Row<size_t>` should
be used instead.
The reason that `arma::Col<size_t>` was originally used was convenience.
Files where `arma::Col<size_t>` is used to hold labels are listed below.
Note that changes in these files will probably mean that the tests will
have to be modified too.
* ./core/data/normalize_labels.hpp
* ./core/data/normalize_labels_impl.hpp
* ./core/tree/cover_tree/cover_tree.hpp
* ./core/tree/cover_tree/cover_tree_impl.hpp
* ./methods/cf/cf.hpp
* ./methods/cf/cf_impl.hpp
* ./methods/cf/cf_main.cpp
* ./methods/det/dtree.cpp
* ./methods/det/dtree.hpp
* ./methods/det/dt_utils.cpp
* ./methods/emst/dtb.hpp
* ./methods/emst/dtb_rules.hpp
* ./methods/emst/dtb_rules_impl.hpp
* ./methods/emst/union_find.hpp
* ./methods/gmm/em_fit.hpp
* ./methods/gmm/em_fit_impl.hpp
* ./methods/gmm/gmm.hpp
* ./methods/gmm/gmm_impl.hpp
* ./methods/hmm/hmm.hpp
* ./methods/hmm/hmm_impl.hpp
* ./methods/hmm/hmm_train_main.cpp
* ./methods/hmm/hmm_viterbi_main.cpp
* ./methods/kmeans/allow_empty_clusters.hpp
* ./methods/kmeans/kmeans.hpp
* ./methods/kmeans/kmeans_impl.hpp
* ./methods/kmeans/kmeans_impl_openmp.hpp
* ./methods/kmeans/kmeans_main.cpp
* ./methods/kmeans/max_variance_new_cluster.hpp
* ./methods/kmeans/max_variance_new_cluster_impl.hpp
* ./methods/kmeans/random_partition.hpp
* ./methods/kmeans/refined_start.hpp
* ./methods/kmeans/refined_start_impl.hpp
* ./methods/naive_bayes/naive_bayes_classifier.hpp
* ./methods/naive_bayes/naive_bayes_classifier_impl.hpp
* ./methods/nca/nca.hpp
* ./methods/nca/nca_impl.hpp
* ./methods/nca/nca_main.cpp
* ./methods/nca/nca_softmax_error_function.hpp
* ./methods/nca/nca_softmax_error_function_impl.hpp
--
--
Ticket URL: <http://trac.research.cc.gatech.edu/fastlab/ticket/350#comment:1>
MLPACK <www.fast-lab.org>
MLPACK is an intuitive, fast, and scalable C++ machine learning library developed at Georgia Tech.
More information about the mlpack-svn
mailing list