confusionmat
Compute confusion matrix for classification problem
Syntax
Description
Examples
Calculate Confusion Matrix
Load a sample of predicted and true labels for a classification problem. trueLabels
are the true labels for an image classification problem and predictedLabels
are the predictions of a convolutional neural network.
load('Cifar10Labels.mat','trueLabels','predictedLabels');
Calculate the numeric confusion matrix. order
is the order of the classes in the confusion matrix.
[m,order] = confusionmat(trueLabels,predictedLabels)
m = 10×10
923 4 21 8 4 1 5 5 23 6
5 972 2 0 0 0 0 1 5 15
26 2 892 30 13 8 17 5 4 3
12 4 32 826 24 48 30 12 5 7
5 1 28 24 898 13 14 14 2 1
7 2 28 111 18 801 13 17 0 3
5 0 16 27 3 4 943 1 1 0
9 1 14 13 22 17 3 915 2 4
37 10 4 4 0 1 2 1 931 10
20 39 3 3 0 0 2 1 9 923
order = 10x1 categorical
airplane
automobile
bird
cat
deer
dog
frog
horse
ship
truck
You can use confusionchart
to plot the confusion matrix as a confusion matrix chart.
figure cm = confusionchart(m,order);
You do not need to calculate the confusion matrix first and then plot it. Instead, plot a confusion matrix chart directly from the true and predicted labels. You can also add column and row summaries and a title.
figure cm = confusionchart(trueLabels,predictedLabels, ... 'Title','My Title', ... 'RowSummary','row-normalized', ... 'ColumnSummary','column-normalized');
The ConfusionMatrixChart
object stores the numeric confusion matrix in the NormalizedValues
property and classes in the ClassLabels
property.
cm.NormalizedValues
ans = 10×10
923 4 21 8 4 1 5 5 23 6
5 972 2 0 0 0 0 1 5 15
26 2 892 30 13 8 17 5 4 3
12 4 32 826 24 48 30 12 5 7
5 1 28 24 898 13 14 14 2 1
7 2 28 111 18 801 13 17 0 3
5 0 16 27 3 4 943 1 1 0
9 1 14 13 22 17 3 915 2 4
37 10 4 4 0 1 2 1 931 10
20 39 3 3 0 0 2 1 9 923
cm.ClassLabels
ans = 10x1 categorical
airplane
automobile
bird
cat
deer
dog
frog
horse
ship
truck
Input Arguments
group
— Known groups
numeric vector | logical vector | character array | string array | cell array of character vectors | categorical vector
Known groups for categorizing observations, specified as a numeric vector, logical vector, character array, string array, cell array of character vectors, or categorical vector.
group
is a grouping variable of the same type as grouphat
. The group
argument must have the same number of observations as grouphat
, as described in Grouping Variables (Statistics and Machine Learning Toolbox). The confusionmat
function treats character arrays and string arrays as cell arrays of character vectors. Additionally, confusionmat
treats NaN
, empty, and 'undefined'
values in group
as missing values and does not count them as distinct groups or categories.
Example: {'Male','Female','Female','Male','Female'}
Data Types: single
| double
| logical
| char
| string
| cell
| categorical
grouphat
— Predicted groups
numeric vector | logical vector | character array | string array | cell array of character vectors | categorical vector
Predicted groups for categorizing observations, specified as a numeric vector, logical vector, character array, string array, cell array of character vectors, or categorical vector.
grouphat
is a grouping variable of the same type as group
. The grouphat
argument must have the same number of observations as group
, as described in Grouping Variables (Statistics and Machine Learning Toolbox). The confusionmat
function treats character arrays and string arrays as cell arrays of character vectors. Additionally, confusionmat
treats NaN
, empty, and 'undefined'
values in grouphat
as missing values and does not count them as distinct groups or categories.
Example: [1 0 0 1 0]
Data Types: single
| double
| logical
| char
| string
| cell
| categorical
grouporder
— Group order
numeric vector | logical vector | character array | string array | cell array of character vectors | categorical vector
Group order, specified as a numeric vector, logical vector, character array, string array, cell array of character vectors, or categorical vector.
grouporder
is a grouping variable containing all the distinct elements in group
and grouphat
. Specify grouporder
to define the order of the rows and columns of C
. If grouporder
contains elements that are not in group
or grouphat
, the corresponding entries in C
are 0
.
By default, the group order depends on the data type of s = [group;grouphat]
:
For numeric vectors, the order is the sorted order of the unique values in
s
.For logical vectors, the order is
false
(0
) followed bytrue
(1
).For categorical vectors, the order is the order returned by
.categories
(s)For other data types, the order is the order of first appearance in
s
.
Example: 'order',{'setosa','versicolor','virginica'}
Data Types: single
| double
| logical
| char
| string
| cell
| categorical
Output Arguments
C
— Confusion matrix
matrix
Confusion matrix, returned as a square matrix with size equal to the total number of distinct elements in the group
and grouphat
arguments. C(i,j)
is the count of observations known to be in group i
but predicted to be in group j
.
The rows and columns of C
have identical ordering of the same group indices. By default, the group order depends on the data type of s = [group;grouphat]
:
For numeric vectors, the order is the sorted order of the unique values in
s
.For logical vectors, the order is
false
(0
) followed bytrue
(1
).For categorical vectors, the order is the order returned by
.categories
(s)For other data types, the order is the order of first appearance in
s
.
To change the order, specify grouporder
,
The confusionmat
function treats NaN
, empty, and 'undefined'
values in the grouping variables as missing values and does not include them in the rows and columns of C
.
order
— Order of rows and columns
numeric vector | logical vector | categorical vector | cell array of character vectors
Order of rows and columns in C
, returned as a numeric vector, logical vector, categorical vector, or cell array of character vectors. If group
and grouphat
are character arrays, string arrays, or cell arrays of character vectors, then the variable order
is a cell array of character vectors. Otherwise, order
is of the same type as group
and grouphat
.
Alternative Functionality
Use
confusionchart
to calculate and plot a confusion matrix. Additionally,confusionchart
displays summary statistics about your data and sorts the classes of the confusion matrix according to the class-wise precision (positive predictive value), class-wise recall (true positive rate), or total number of correctly classified observations.
Version History
R2024b: Matrix returned when you use one logical observation
When you create a confusion matrix using only one observation of type
logical
, the confusionmat
function returns a
matrix instead of a scalar value. The default order of the rows and columns is
false
(0
) followed by true
(1
).
confusionmat(true,true)
ans = 0 0 0 1
MATLAB Command
You clicked a link that corresponds to this MATLAB command:
Run the command by entering it in the MATLAB Command Window. Web browsers do not support MATLAB commands.
Select a Web Site
Choose a web site to get translated content where available and see local events and offers. Based on your location, we recommend that you select: .
You can also select a web site from the following list:
How to Get Best Site Performance
Select the China site (in Chinese or English) for best site performance. Other MathWorks country sites are not optimized for visits from your location.
Americas
- América Latina (Español)
- Canada (English)
- United States (English)
Europe
- Belgium (English)
- Denmark (English)
- Deutschland (Deutsch)
- España (Español)
- Finland (English)
- France (Français)
- Ireland (English)
- Italia (Italiano)
- Luxembourg (English)
- Netherlands (English)
- Norway (English)
- Österreich (Deutsch)
- Portugal (English)
- Sweden (English)
- Switzerland
- United Kingdom (English)