Categories &

Functions List

Class Definition: ConfusionMatrixChart

statistics: ConfusionMatrixChart

Confusion matrix chart for classification results

The ConfusionMatrixChart class implements a confusion matrix chart object, which displays the classification performance of a classifier by showing the counts of true positive, true negative, false positive, and false negative predictions.

A confusion matrix chart is a visual representation of the performance of a classification algorithm. The rows represent the true classes and the columns represent the predicted classes. The diagonal elements represent the correctly classified observations, while the off-diagonal elements represent the misclassified observations.

Create a ConfusionMatrixChart object by using the confusionchart function.

See also: confusionchart

Source Code: ConfusionMatrixChart

Properties

A character vector specifying the label for the x-axis. Default is "Predicted Class".

A character vector specifying the label for the y-axis. Default is "True Class".

A character vector specifying the title of the confusion matrix chart. Default is empty string.

A character vector specifying the font name used for all text elements in the chart. Default is empty string, which uses the axes font name.

A numeric scalar specifying the font size used for all text elements in the chart. Default is 0, which uses the axes font size.

A 1x3 RGB vector specifying the color for the diagonal elements of the confusion matrix, which represent correct classifications. Default is [0.0, 0.4471, 0.7412].

A 1x3 RGB vector specifying the color for the off-diagonal elements of the confusion matrix, which represent misclassifications. Default is [0.8510, 0.3255, 0.0980].

A character vector specifying how to normalize the confusion matrix values. Supported values are:

  • "absolute" - Display absolute counts (default)
  • "column-normalized" - Normalize by column totals
  • "row-normalized" - Normalize by row totals
  • "total-normalized" - Normalize by total number of observations

A character vector specifying whether and how to display column summaries. Supported values are:

  • "off" - Do not display column summary (default)
  • "absolute" - Display absolute counts
  • "column-normalized" - Display normalized by column
  • "total-normalized" - Display normalized by total

A character vector specifying whether and how to display row summaries. Supported values are:

  • "off" - Do not display row summary (default)
  • "absolute" - Display absolute counts
  • "row-normalized" - Display normalized by row
  • "total-normalized" - Display normalized by total

A character vector specifying whether to display grid lines in the confusion matrix. Supported values are:

  • "on" - Display grid lines (default)
  • "off" - Hide grid lines

A character vector specifying the visibility of the object’s handle. Supported values are "on", "off", and "callback".

A 1x4 numeric vector specifying the outer position of the chart in the format [left, bottom, width, height].

A 1x4 numeric vector specifying the position of the chart in the format [left, bottom, width, height].

A character vector specifying the units for the position properties. Supported values are "centimeters", "characters", "inches", "normalized", "pixels", and "points".

A cell array of character vectors containing the class labels used in the confusion matrix. This property is read-only.

A numeric matrix containing the normalized confusion matrix values according to the current normalization setting. This property is read-only.

A handle to the parent figure or container object. This property is read-only.

Methods

statistics: cmc = ConfusionMatrixChart (hax, cm, cl)
statistics: cmc = ConfusionMatrixChart (…, name, value)

cmc = ConfusionMatrixChart (hax, cm, cl) returns a ConfusionMatrixChart object with parent axes hax, confusion matrix cm, and class labels cl.

  • hax must be a valid axes handle where the chart will be displayed.
  • cm must be a square numeric matrix containing the confusion matrix values, where rows represent true classes and columns represent predicted classes.
  • cl must be a cell array of character vectors containing the class labels. The number of labels must match the size of the confusion matrix.

cmc = ConfusionMatrixChart (…, name, value) returns a ConfusionMatrixChart object with additional parameters specified by name, value paired arguments:

NameValue
"XLabel"A character vector specifying the x-axis label. Default is "Predicted Class".
"YLabel"A character vector specifying the y-axis label. Default is "True Class".
"Title"A character vector specifying the chart title. Default is empty string.
"FontName"A character vector specifying the font name for text elements. Default is the axes font name.
"FontSize"A numeric scalar specifying the font size for text elements. Default is the axes font size.
"DiagonalColor"A 1x3 RGB vector specifying the color for diagonal elements (correct classifications). Default is [0.0, 0.4471, 0.7412].
"OffDiagonalColor"A 1x3 RGB vector specifying the color for off-diagonal elements (misclassifications). Default is [0.8510, 0.3255, 0.0980].
"Normalization"A character vector specifying the normalization method. Supported values are "absolute", "column-normalized", "row-normalized", and "total-normalized". Default is "absolute".
"ColumnSummary"A character vector specifying whether and how to display column summaries. Supported values are "off", "absolute", "column-normalized", and "total-normalized". Default is "off".
"RowSummary"A character vector specifying whether and how to display row summaries. Supported values are "off", "absolute", "row-normalized", and "total-normalized". Default is "off".
"GridVisible"A character vector specifying whether to display grid lines. Supported values are "on" and "off". Default is "on".
"HandleVisibility"A character vector specifying the handle visibility. Supported values are "on", "off", and "callback".
"OuterPosition"A 1x4 numeric vector specifying the outer position of the chart.
"Position"A 1x4 numeric vector specifying the position of the chart.
"Units"A character vector specifying the position units. Supported values are "centimeters", "characters", "inches", "normalized", "pixels", and "points".

See also: confusionchart

ConfusionMatrixChart: disp (cmc)

disp (cmc) displays the main properties of the ConfusionMatrixChart object cmc, including the normalized values and class labels.

See also: ConfusionMatrixChart

ConfusionMatrixChart: sortClasses (cmc, order)

sortClasses (cmc, order) sorts the classes in the confusion matrix chart cmc according to the specified order.

order can be:

  • A cell array of class labels in the desired order
  • "auto" - Sort class labels alphabetically
  • "ascending-diagonal" - Sort by ascending diagonal values
  • "descending-diagonal" - Sort by descending diagonal values
  • "cluster" - Sort using hierarchical clustering

When using "cluster", the classes are grouped based on similarity using hierarchical clustering, which can help identify groups of frequently confused classes.

See also: confusionchart, linkage, pdist

Examples

  
 cm = ConfusionMatrixChart (gca, [1 2; 1 2], {"A","B"}, {"XLabel","LABEL A"})
 NormalizedValues = cm.NormalizedValues
 ClassLabels = cm.ClassLabels
  
cm =

ConfusionMatrixChart with properties:

	NormalizedValues: [ 2x2 double ]
	ClassLabels: { 1x2 cell }


NormalizedValues =

   1   2
   1   2

ClassLabels =
  1x2 cell array

    {'A'}    {'B'}
plotted figure