ndhistc

Multi dimensional histogram

function qHistogram = ndHistc (mData, vEdge1, vEdge2, ... )

* Input Arguments:

+ mData: nRecord by nDim 2-dimensional array of doubles
+ vEdge1, vEdge2, ... : nDim vectors of histogram edges

* Return value:

+ qHistogram: nDim-dimensional data cube
containing number of points in each cell
defined by histogram edges. For instance,
qHistogram(1,1,1,...) means number of data points
satisfying
vEdge1(1) <= mData(:,1) < vEdge1(2) & ...
vEdge2(1) <= mData(:,2) < vEdge2(2) & ...
vEdge3(1) <= mData(:,3) < vEdge3(1) & ...
...
* Example
mRand = rand(1e6,5);
ve1 = linspace(0,1,5);
ve2 = linspace(0,1,6);
ve3 = linspace(0,1,7);
ve4 = linspace(0,1,8);
ve5 = linspace(0,1,9);
qHist = ndhistc(mRand, ve1, ve2, ve3, ve4, ve5);

* Comparison with ndhist.m (compiled using mcc -x ndhist)
+ 1e6 by 2 data -> 5 by 6
ndhist.m 79.49 sec
ndhistc.c 0.4610 sec

+ 1e6 by 5 data -> 5 by 6 by 7 by 8 by 9
ndhist.m 199.32 sec
ndhistc.c 2.4430 sec

==> More efficient if More data points (==rows) & Less dimensions (==columns)