⚠️There's a newer version (1.0.3) of this package.Take me there.

ldamatch (version 0.6.3)

Multivariate Condition Matching by Backwards Elimination Using Linear Discriminant Analysis

Description

Performs group matching by backward elimination using linear discriminant analysis.

Copy Link

Version

Install

install.packages('ldamatch')

Monthly Downloads

245

Version

0.6.3

License

MIT + file LICENSE

Maintainer

Kyle Gorman

Last Published

August 26th, 2015

Functions in ldamatch (0.6.3)

create_Cartesian_iterable

Creates Cartesian product of iterators.

search_exhaustive

Searches the space backwards, prefering more subjects and better ratios. While the search is done in parallel, the search space is enormous and so it can be very slow in the worst case. It is perhaps most useful as a tool to study other matching procedures.

check_subspaces_for_group_size_setup

Searches over all possible subspaces for specified group size setup.

apply_crit

Returns smallest halting_test() / thresh ratio, or zero if less than 1.

t_halt

A univariate halting test using the t-test (t.test).

U_halt

A univariate halting test using the Wilcoxon test (wilcox.test).

create_subject_subspace_using_LDA

Creates ordered subspace of subject candidates for removal, using LDA.

get_param

Gets parameter value for ldamatch.

search_montecarlo

Searches for a subset satisfying the test by randomly selecting subspaces of decreasing size.

ks_crit

Criterion function for ks_halt.

combine_ratios

Combines matched group candidate with current best one.

ad_crit

Criterion function for ad_halt.

t_crit

Criterion function for ad_halt.

decrease_group_sizes

Creates all group sizes by reducing one group in all rows of grpsizes.

search_heuristic

At each vertex of the search graph, this takes a step which moves the proportions of conditions in the subspace closer to the desired (or sample) proportions.

GUtils

ldamatch: Multivariate Condition Matching by Backwards Elimination using Fisher's Linear Discriminant

ks_halt

A univariate halting test using the Kolmogorov-Smirnov Test (ks.test).

wilks_halt

A multivariate halting test (appropriate when nlevels(condition) > 2)).

U_crit

Criterion function for U_halt.

create_halting_test

Creates halt test from multiple tests.

get_human_readable

Returns human readable format for number of seconds.

estimate_exhaustive

Estimates the maximum number of cases to be checked during exhaustive search.

ad_halt

A univariate halting test using the Anderson-Darling test (kSamples::ad.test).

ldamatch

Creates a matched group via backward selection.

set_param

Sets parameters for ldamatch.