butools.trace.PdfFromTrace¶

butools.trace.PdfFromTrace()¶

Matlab:	[x, y] = PdfFromTrace(trace, intBounds)
Mathematica:	{x, y} = PdfFromTrace[trace, intBounds]
Python/Numpy:	x, y = PdfFromTrace(trace, intBounds)

Returns the empirical density function of a trace.

Parameters:	trace : vector of doubles The trace data intBounds : vector of doubles The array of interval boundaries. The pdf is the number of samples falling into an interval divided by the interval length.
Returns:	x : vector of doubles The center of the intervals (the points where the empirical pdf is calculated) y : vector of doubles The values of the empirical pdf at the given points

Examples

For Matlab:

>>> D0 = [-18., 1., 4.; 2., -18., 7.; 1., 3., -32.];
>>> D1 = [12., 1., 0.; 1., 8., 0.; 2., 1., 25.];
>>> x = (0.0:0.01:0.5);
>>> tr = SamplesFromMAP(D0, D1, 1000000);
>>> [x, y] = PdfFromTrace(tr, x);
>>> [a, A] = MarginalDistributionFromMAP(D0, D1);
>>> [xm, ym] = IntervalPdfFromPH(a, A, x);
>>> plot(x, y, xm, ym)

For Mathematica:

>>> D0 = {{-18., 1., 4.},{2., -18., 7.},{1., 3., -32.}};
>>> D1 = {{12., 1., 0.},{1., 8., 0.},{2., 1., 25.}};
>>> x = Range[0.0,0.5,0.01];
>>> tr = SamplesFromMAP[D0, D1, 1000000];
>>> {x, y} = PdfFromTrace[tr, x];
>>> {a, A} = MarginalDistributionFromMAP[D0, D1];
>>> {xm, ym} = IntervalPdfFromPH[a, A, x];
>>> ListLinePlot[{Transpose[{x, y}],Transpose[{xm, ym}]}]

For Python/Numpy:

>>> D0 = ml.matrix([[-18., 1., 4.],[2., -18., 7.],[1., 3., -32.]])
>>> D1 = ml.matrix([[12., 1., 0.],[1., 8., 0.],[2., 1., 25.]])
>>> x = np.arange(0.0,0.51,0.01)
>>> tr = SamplesFromMAP(D0, D1, 1000000)
>>> x, y = PdfFromTrace(tr, x)
>>> a, A = MarginalDistributionFromMAP(D0, D1)
>>> xm, ym = IntervalPdfFromPH(a, A, x)
>>> plt.plot(x, y, xm, ym)