topsClassification Class Reference

Overview

Represents data as points in space, for classification.

topsClassification takes in data samples from multiple sources and outputs arbitrary values. It maps samples to outputs by way of a spatial model: each data source maps to a spatial dimension, each sample maps to a value along one dimension, and a set of samples maps to a point in space. Regions in the space map to arbitrary outputs, so the region in which a point falls determines the output for a sample set.

The spatial model is backed up by a high-dimensional matrix. Picking a point in space boils down to indexing into the matrix. The key is to convert arbitrary data samples into matrix indices.

Inheritance diagram for topsClassification:

Properties and Events
Property	defaultOutput = []
	default output when no other output is defined
Property	defaultOutputName = 'default'
	name to give to the default output
Properties and Events inherited from topsFoundation
Property	name = ''
	a string name to indentify the object
Property	clockFunction =@topsClock
	clock function, to standardize

Protected Properties
	sources
	struct array of data sources and descriptions
Property	space
	topsSpace with spatial modeling utilities
	outputs
	struct array of spatial regions and outputs
Property	outputTable
	high-dimensional matrix for looking up output

Methods
self	topsClassification (varargin)
	Constuct with name optional. More...
fig	gui (self)
	Open a GUI to view object details. More...
index	addSource (self, name, sampleFunction, minimum, maximum, nPoints)
	Add a data source. More...
index	removeSource (self, name)
	Remove a data source by name. More...
index	addOutput (self, name, region, value)
	Add a classification output. More...
index	removeOutput (self, name)
	Remove a classification output by name. More...
index	editOutputValue (self, name, value)
	Change the value mapped from an existing output. More...
	updateSamples (self)
	Update source samples. More...
value name	getOutput (self, doUpdate)
	Get the current classification output. More...
Methods inherited from topsFoundation
self	topsFoundation (name)
	Constuct with name optional. More...
fig	gui (self)
	Open a GUI to view object details. More...
p	guiPanel (self, varargin)
	Make a topsDetailPanel with details about this object. More...

Protected Methods
	buildSpace (self)
	Rebuild the data space. More...
	clearOutputs (self)
	Clear out the outputTable lookuptable.
	buildOutputTable (self)
	Rebuild the outputTable lookuptable.

Additional Inherited Members
Static Methods inherited from topsFoundation
static c index	cellAdd (c, item, index)
	Add an item to a cell array. More...
static selector	cellContains (c, item)
	Does a cell array conatin an item? More...
static c	cellRemoveItem (c, item)
	Remove an item from a cell array. More...
static c	cellRemoveElement (c, index)
	Remove indexed elements from a cell array. More...
static index selector	findStructName (s, name)
	Where is the given name in the given struct array? More...

Constructor

topsClassification()

self topsClassification

(

varargin

)

Constuct with name optional.

Parameters

name	optional name for this object

If name is provided, assigns name to this object.

Methods

gui()

fig gui

(

self

)

Open a GUI to view object details.

Opens a new GUI with components suitable for viewing objects of this class. Returns a topsFigure object which contains the GUI.

addSource()

index addSource	(	self	,
		name	,
		sampleFunction	,
		minimum	,
		maximum	,
		nPoints
	)

Add a data source.

Parameters

name	unique name for the data source
sampleFunction	function_handle that returns a sample
minimum	minimum sample value to expect
maximum	maximum sample value to expect
nPoints	points to use between minumum and maximum

Adds a data source to be included in classification. name must be unique so that the source can be referred to later by name. Any existing source with the same name will be replaced. This will invalidate existing outputs, because it changes the spatial model.

sampleFunction must be function handle, and the function should return a scalar, numeric sample value for the named source.

minimum and maximum must be the minumum and maximum expected sample values from the named source. nPoints must be the number of points to space evenly between minimum and maximum. minimum, maximum, and nPoints are used to convert raw sample values into high-dimensional matrix indices.

Returns the intex into sources where the source was inserted or appended.

removeSource()

index removeSource	(	self	,
		name
	)

Remove a data source by name.

Parameters

name	unique name of a data source

Removes the named data source. This will invalidate existing outputs, because it changes the spatial model.

Returns the index into sources where the named source was removed, or [] if no source was found.

addOutput()

index addOutput	(	self	,
		name	,
		region	,
		value
	)

Add a classification output.

Parameters

name	unique name for the output
region	topsRegion spatial region that maps to the output
value	output value mapped from region

Adds an output value that can be returned as a classification result. region must be a topsRegion object specifying the region of the spatial model that should map to value. name must be a unique name for this region-value pair.

If multiple output regions overlap, those that were added later take presidence.

Returns the index into outputs where the named output was inserted or appended.

removeOutput()

index removeOutput	(	self	,
		name
	)

Remove a classification output by name.

Parameters

name	unique name of a classification output

Removes the named output from classification. Returns the index into outputs where the named output was removed, or [] if no output was found.

editOutputValue()

index editOutputValue	(	self	,
		name	,
		value
	)

Change the value mapped from an existing output.

Parameters

name	unique name of a classification output
value	new value for the named output

Change the output value associated with the given name. The spatial region associated with name remians the same, but the value, as returned from getOutput() is replaced with the new value.

Returns the index into outputs where the named value was changed, or [] if no output was found.

updateSamples()

updateSamples

(

self

)

Update source samples.

Retrieves new samples for each data source, using the sampleFunction provided for each source in addSource().

getOutput()

value name getOutput	(	self	,
		doUpdate
	)

Get the current classification output.

Parameters

doUpdate

optional, whether invoke updateSamples() first

Uses the current sample value of all sources to locate a point in the spatial model and return the associated output value. Returns as a second output the name of the output. By default, invokes updateSamples() to refresh source sample values. if doUpdate is provided and false, leaves source samples as they are.

buildSpace()

buildSpace ( self  )

protected

Rebuild the data space.

Invalidate now-stale outputs and reallocate the outputTable.

---

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

• topsClassification.m