# ITE / code / shared / embedded / KDP / mat_oct / kdpee.m

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40``` ```function h = kdpee (arr, mins, maxs, zcut) % function h = kdpee (arr) % function h = kdpee (arr, mins, maxs) % function h = kdpee (arr, mins, maxs, zcut) % % Entropy estimator using k-d partitioning. Returns a value in nats. % % "arr" is a matrix of data points, 1 row per datum. % % The algorithm can run on data with known OR unknown support. % (Support is assumed to be extent of data in the latter case.) % If the support is known, pass it in as vectors "mins" and "maxs". % If the support is not known simply don't supply "mins" or "maxs". % % kdpee is free software: you can redistribute it and/or modify % it under the terms of the GNU General Public License as published by % the Free Software Foundation, either version 3 of the License, or % (at your option) any later version. if (nargin < 4) zcut = 1.96; end n = size(arr, 1); % num data points d = size(arr, 2); % dimensionality %if (n < (2^d)) % fprintf(1, 'kdpee() warning: n < 2^d, meaning you might not have enough data to make an entropy estimate for the dimensionality\n'); %end if(nargin < 3) % Take limits from data, should have tendency to underestimate support % but seems to work well mins = min(arr, [], 1); maxs = max(arr, [], 1); end h = kdpeemex(arr, mins, maxs, zcut); ```