Postprocess.StochasticAssessment.Output Class Reference

Detailed Description

result class for computations executed by an Postprocess.StochasticAssessment.Assessment object.

This class provides methods to transform and visualize the gathered data for publication in LaTeX/Tikz documents.

Definition at line 19 of file Output.m.

Inheritance diagram for Postprocess.StochasticAssessment.Output:

Collaboration diagram for Postprocess.StochasticAssessment.Output:

Public Member Functions
	Output (values,Postprocess.StochasticAssessment.Assessment assessment)
	Constructor. More...
function	plot_landscape ()
	plots a two dimensional error landscape in case of two dimensional plot_fields More...
function	print_table (datafile)
	prints out a CSV table with gathered data. More...
function	print_landscape (datafile, fieldname)
	equivalent to the plot_landscape() function put prints out a table which can be interpreted e.g. by pgfplots/Tikz. More...
function	print_3d_curve (datafile, fieldname)
	equivalent to the print_landscape() function put prints out a table which can be interpreted for a 3d-curve plot e.g. by pgfplots/Tikz. More...
function slic =	get_slice (fieldname, cell_range, struct_range)
	helper function slicing a field in the values attribute More...
function rslice =	reduce_slice (red_func, fieldname, cell_range, struct_range, direction)
	extends the get_slice() function by also applying a reduction function on each matrix entry. More...
function	merge (Postprocess.StochasticAssessment.Output other)
	helper function merging this Output object with another one. More...

Public Attributes
	values
	values computed by Postprocess.StochasticAssessment.Assessment More...
	rd_conds
	cell array of all condition numbers of reduced simulaton system matrices.
	rtime_est = {""}
	time for the computation of the error estimator. By default this is not set. More...
Assessment	assessment
	the object with which the values are generated More...
	error_bound = 1
	maxmimum error values to be plotted. Above this values error display is cropped. More...
	estimator_bound = 100
	maxmimum estimator values to be plotted. Above this values error display is cropped. More...
	stab_limit = 1e-2
	define stability-region as error being smaller than sqrt(diffmax * area), e.g. diffmax = 4, area = 2e-7 stability region More...

Constructor & Destructor Documentation

Postprocess.StochasticAssessment.Output.Output	(		values,
		Postprocess.StochasticAssessment.Assessment	assessment
	)

Constructor.

Parameters

values	the structure with the gathered data by the Postprocess.StochasticAssessment.Assessment object
assessment	the generating object

Definition at line 116 of file Output.m.

Member Function Documentation

function slic = Postprocess.StochasticAssessment.Output.get_slice	(		fieldname,
			cell_range,
			struct_range
	)

helper function slicing a field in the values attribute

Parameters

fieldname	name of the field in the values attributed to be sliced
cell_range	cell indices to be extracted from the field. If empty all cell entries are returned Default: []
struct_range	Usually the values attribute is a cell of structure arrays. arrays. This range defines a slice in the structure array. If empty all array entries are returned. Default: []

Return values

slic	the sliced values attribute

Definition at line 431 of file Output.m.

Here is the caller graph for this function:

function Postprocess.StochasticAssessment.Output.merge

(

Postprocess.StochasticAssessment.Output

other

)

helper function merging this Output object with another one.

The following rules apply during merging

1. If values.max_Delta is all Inf, copy max_Delta and max_Delta_ind from other.
2. If values.max_err is all Inf, copy max_err and max_err_ind from other.
3. If one of rconds, dconds or rtime_est is empty, copy it from other
4. If one of rtime, rrtime or dtime is set to all-NaN, copy it from other
5. If this and other have valid rtime (i.e. not-Nan) and only one of both has computed estimators, compute rtime_est by subtracting one from the other.

Parameters

other

another object to be merged into this one.

Required fields of other:

• values —  values
• assessment.compute_estimates —  assessment.compute estimates
• rtime_est —  rtime est
• rd_conds —  rd conds

Definition at line 508 of file Output.m.

function Postprocess.StochasticAssessment.Output.plot_landscape

(

)

plots a two dimensional error landscape in case of two dimensional plot_fields

Used attributes

• error_bound and
• stab_limit

Definition at line 133 of file Output.m.

Here is the call graph for this function:

function Postprocess.StochasticAssessment.Output.print_3d_curve	(		datafile,
			fieldname
	)

equivalent to the print_landscape() function put prints out a table which can be interpreted for a 3d-curve plot e.g. by pgfplots/Tikz.

This function can be used to display certain curves on the error landscape generated by print_landscape().

An example for a pgfplots graphics that can interpret the table output is shown in the following code snippet:

\begin{axis}[zmode=log,log basis z=10,axis on top,
             view={-50}{-30}, axis lines=box,
             xlabel={$N$}, ylabel={$M$},
             zlabel={$\eta_{N,M,M'}$}]
 \addplot3[mark=0]
 table[x=N, y=M, z=error] {error_curve.dat};

Parameters

datafile	a string specifying the path and filename of the data table to be stored
fieldname	the fieldname in the values attribute that shall be plotted Default: max_err

Definition at line 382 of file Output.m.

Here is the call graph for this function:

function Postprocess.StochasticAssessment.Output.print_landscape	(		datafile,
			fieldname
	)

equivalent to the plot_landscape() function put prints out a table which can be interpreted e.g. by pgfplots/Tikz.

An example for a pgfplots graphics that can interpret the table output is shown in the following code snippet:

\begin{axis}[zmode=log,log basis z=10,axis on top,
             view={-50}{-30}, axis lines=box,
             xlabel={$N$}, ylabel={$M$},
             zlabel={$\eta_{N,M,M'}$}]
 \addplot3[surf]
 table[x=N, y=M, z=error] {error_landscape.dat};

Parameters

datafile	a string specifying the path and filename of the data table to be stored
fieldname	the fieldname in the values attribute that shall be plotted Default: max_err

Definition at line 340 of file Output.m.

Here is the call graph for this function:

function Postprocess.StochasticAssessment.Output.print_table

(

datafile

)

prints out a CSV table with gathered data.

An example for a pgfplotstable graphic that can interpret the table output is shown in the following code snippet:

\pgfplotstabletypeset[
  font={\footnotesize},
  columns={dim1,dim2,{time_avg},{err_max},{offtime_total}},
  columns/{dim1}/.style={
    column name={N},

      \ifnum\pgfplotstablerow=0

      \fi

  },
  columns/{dim2}/.style={
    column name={M},

      \ifnum\pgfplotstablerow=0

      \fi
    },
  },
  columns/{time_avg}/.style={
    column name={\o-run--time[s]}
  },
  columns/{err_max}/.style={
    column name={max. error}
  },
  columns/{offtime_total}/.style={
    column name={offline time[h]},
    divide by=3600
  },
  every even row/.style={
    before row={\rowcolor[gray]{0.9}}},
  every head row/.style={
    before row=\toprule,after row=\midrule},
  every last row/.style={
    after row=\bottomrule},
  display columns/2/.style={dec sep align},
  display columns/3/.style={zerofill,precision=2},
  empty cells with={\ensuremath{-}}
]{table.dat};

This function prints out a table with data for the average, min and max values of

• reduced simulation time,
• reduced simulation reconstruction time,
• reduced simulation error estimator computation time,
• maximum error estimator and the
• maximum "true" error Furthermore, the offline time is printed out split up into
• the total time spent in the offline phase,
• the time spent for (pre-)computation of detailed simulations
• the time spent in the EI-greedy algorithm
• the time spent in the reduced basis generation algorithm

Parameters

datafile

a string specifying the path and filename of the data table to be stored

Definition at line 196 of file Output.m.

Here is the call graph for this function:

function rslice = Postprocess.StochasticAssessment.Output.reduce_slice	(		red_func,
			fieldname,
			cell_range,
			struct_range,
			direction
	)

extends the get_slice() function by also applying a reduction function on each matrix entry.

Parameters

red_func	function pointer who is applied to each matrix entry in the slice, e.g. '@max or @mean'
fieldname	name of the field in the values attributed to be sliced
cell_range	cell indices to be extracted from the field. If empty all cell entries are returned. Default: []
struct_range	Usually the values attribute is a cell of structure arrays. arrays. This range defines a slice in the structure array. If empty all array entries are returned. Default: []
direction	integer specifying in which direction a matrix shall be reduced. In case it is equal to 2, the matrix is transposed before red_func is applied. Default: 1

Return values

rslice

the sliced and reduced values attribute

Definition at line 462 of file Output.m.

Here is the call graph for this function:

Here is the caller graph for this function:

Member Data Documentation

Postprocess.StochasticAssessment.Output.assessment

the object with which the values are generated

Todo:

Maybe we should cache this, too...

Definition at line 64 of file Output.m.

Postprocess.StochasticAssessment.Output.error_bound = 1

maxmimum error values to be plotted. Above this values error display is cropped.

Default: = 1;

Definition at line 74 of file Output.m.

Postprocess.StochasticAssessment.Output.estimator_bound = 100

maxmimum estimator values to be plotted. Above this values error display is cropped.

Default: = 100;

Definition at line 85 of file Output.m.

Postprocess.StochasticAssessment.Output.rtime_est = {""}

time for the computation of the error estimator. By default this is not set.

Default: {""}

Definition at line 56 of file Output.m.

Postprocess.StochasticAssessment.Output.stab_limit = 1e-2

define stability-region as error being smaller than sqrt(diffmax * area), e.g. diffmax = 4, area = 2e-7 stability region

If the averaged error exceeds this value, the visualization is cropped here. This happens especially for small reduced basis sizes. (Default = 1e-2)

Default: 1e-2

Definition at line 97 of file Output.m.

Postprocess.StochasticAssessment.Output.values

values computed by Postprocess.StochasticAssessment.Assessment

Usually this is a cell array of structure arrays. For cells correspond to the rsamples and the structure array corresponds to the test samples given by M_test.

Definition at line 36 of file Output.m.

---

The documentation for this class was generated from the following file:

• general/+Postprocess/+StochasticAssessment/Output.m