COMSOL 6.2 - KernelDensityEstimation

Compute a kernel density estimation (KDE) from a table of values.

model.result().dataset().create(<dtag>,"KernelDensityEstimation");

model.result().dataset(<dtag>).set(property,<value>);

model.result().dataset(<dtag>).setIndex(property,<value>,<index>);

model.result().dataset().create(<dtag>,"KernelDensityEstimation"); creates a kernel density evaluation dataset.

Kernel density estimation datasets take a table and compute the kernel density estimation (KDE), which is a nonparametric way to estimate the probability density function of a random variable.

The following properties are available:

Table 7-70: Valid Property/Value Pairs for KernelDensityEstimation Datasets.
Name	Value	Default	Description
autores	Boolean	true	Use an automatic or manual resolution for the variables.
bandwidth	auto \| manual	auto	Use automatic computation of the bandwidth using Silverman’s rule of thumb or provide the bandwidth manually.
bandwidth1d	scalar double	1	The bandwidth for the KDE kernel, if bandwidth is set to manual and dimension is 1.
bandwidth2d	double 2-by-2 array	{1 0; 0 1}	The bandwidth for the KDE kernel, if bandwidth is set to manual and dimension is 2. Specify using setIndex syntax.
bandwidth3d	double 3-by-3 array	{1 0 0; 0 1 0; 0 0 1}	The bandwidth for the KDE kernel, if bandwidth is set to manual and dimension is 3. Specify using setIndex syntax
dimension	1, 2 or 3	1	The dimension of the input data.
evaluationgroup	tag of evaluation group \| none		The evaluation group to use as input data, when source is set to evaluationgroup.
kerneltype	gaussian \| parabolic	gaussian	The type of kernel: Gaussian (normal) or parabolic (Epanechnikov).
outvar	String		Name of the output variable with the KDE value.
resx	positive integer	10	Resolution for the x variable, if autores is true.
resy	positive integer	10	Resolution for the y variable, if autores is true and dimension is 2 or 3.
resz	positive integer	10	Resolution for the zvariable, if autores is true and dimension is 3.
source	table \| evaluationgroup	table	The data source: a table or an evaluation group.
table	tag of table \| none		The table to use as input data, when source is set to table.
xvar	String		Name of input variable for x.
yvar	String		Name of input variable for y, if dimension is 2 or 3.
zvar	String		Name of input variable for z, if dimension is 3.