Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Re-implement stat_surrogate_pvals #299

Merged
merged 3 commits into from
May 17, 2021
Merged

Re-implement stat_surrogate_pvals #299

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
56b3db2
Re-implement stat_surrogate_pvals with an O(n) algorithm and support …
tstenner May 17, 2021
90eec09
Compatibility with Matlab < R2016b
tstenner May 17, 2021
91e761a
Handle observed==distribution for left- and two-tailed tests
tstenner May 17, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
45 changes: 19 additions & 26 deletions functions/statistics/stat_surrogate_pvals.m
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,36 +52,29 @@
% THE POSSIBILITY OF SUCH DAMAGE.

function pvals = stat_surrogate_pvals(distribution,observed,tail)
numDims = ndims(distribution);
if iscolumn(distribution)
numDims = 1;
end

numDims = myndims(distribution);

% append observed to last dimension of surrogate distribution
distribution = cat(numDims,distribution,observed);

numDims = myndims(distribution);

% sort along last dimension (replications)
[tmp idx] = sort( distribution, numDims,'ascend');
[tmp mx] = max( idx,[], numDims);
n = size(distribution, numDims);
% once support for matlab <= R2016b is dropped:
% pvals = sum(distribution >= observed, numDims) / n;
pvals = sum(bsxfun(@ge, distribution, observed), numDims) / n;

len = size(distribution, numDims );
pvals = 1-(mx-0.5)/len;
if strcmpi(tail, 'both')
pvals = min(pvals, 1-pvals);
pvals = 2*pvals;
if any(strcmpi(tail, {'right', 'one'}))
% nothing to be done
return;
end

p_left = 1 - pvals + sum(bsxfun(@eq, distribution, observed), numDims) / n;

% get the number of dimensions in a matrix
function val = myndims(a)
if ndims(a) > 2
val = ndims(a);
if strcmpi(tail, 'both')
pvals = 2 * min(pvals, p_left);
elseif strcmpi(tail, 'left')
pvals = p_left;
else
if size(a,1) == 1,
val = 2;
elseif size(a,2) == 1,
val = 1;
else
val = 2;
end
error('invalid value for tail: "%s", should be left, right, one or both', tail);
end

end