# confusionmat

Compute confusion matrix for classification problem

## Syntax

## Description

uses `C`

= confusionmat(`group`

,`grouphat`

,`'Order'`

,`grouporder`

)`grouporder`

to order the rows and columns of `C`

.

## Examples

### Display Confusion Matrix

Display the confusion matrix for data with two misclassifications and one missing classification.

Create vectors for the known groups and the predicted groups.

g1 = [3 2 2 3 1 1]'; % Known groups g2 = [4 2 3 NaN 1 1]'; % Predicted groups

Return the confusion matrix.

C = confusionmat(g1,g2)

`C = `*4×4*
2 0 0 0
0 1 1 0
0 0 0 1
0 0 0 0

The indices of the rows and columns of the confusion matrix `C`

are identical and arranged by default in the sorted order of `[g1;g2]`

, that is, `(1,2,3,4)`

.

The confusion matrix shows that the two data points known to be in group 1 are classified correctly. For group 2, one of the data points is misclassified into group 3. Also, one of the data points known to be in group 3 is misclassified into group 4. `confusionmat`

treats the `NaN`

value in the grouping variable `g2`

as a missing value and does not include it in the rows and columns of `C`

.

Plot the confusion matrix as a confusion matrix chart by using `confusionchart`

.

confusionchart(C)

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 by using `confusionchart`

.

cm = confusionchart(g1,g2)

cm = ConfusionMatrixChart with properties: NormalizedValues: [4x4 double] ClassLabels: [4x1 double] Show all properties

The `ConfusionMatrixChart`

object stores the numeric confusion matrix in the `NormalizedValues`

property and the classes in the `ClassLabels`

property. Display these properties using dot notation.

cm.NormalizedValues

`ans = `*4×4*
2 0 0 0
0 1 1 0
0 0 0 1
0 0 0 0

cm.ClassLabels

`ans = `*4×1*
1
2
3
4

### Specify Group Order of Confusion Matrix

Display the confusion matrix for data with two misclassifications and one missing classification, and specify the group order.

Create vectors for the known groups and the predicted groups.

g1 = [3 2 2 3 1 1]'; % Known groups g2 = [4 2 3 NaN 1 1]'; % Predicted groups

Specify the group order and return the confusion matrix.

`C = confusionmat(g1,g2,'Order',[4 3 2 1])`

`C = `*4×4*
0 0 0 0
1 0 0 0
0 1 1 0
0 0 0 2

The indices of the rows and columns of the confusion matrix `C`

are identical and arranged in the order specified by the group order, that is, `(4,3,2,1)`

.

The second row of the confusion matrix `C`

shows that one of the data points known to be in group 3 is misclassified into group 4. The third row of `C`

shows that one of the data points belonging to group 2 is misclassified into group 3, and the fourth row shows that the two data points known to be in group 1 are classified correctly. `confusionmat`

treats the `NaN`

value in the grouping variable `g2`

as a missing value and does not include it in the rows and columns of `C`

.

### Confusion Matrix for Classification

Perform classification on a sample of the `fisheriris`

data set and display the confusion matrix for the resulting classification.

Load Fisher's iris data set.

`load fisheriris`

Randomize the measurements and groups in the data.

rng(0,'twister'); % For reproducibility numObs = length(species); p = randperm(numObs); meas = meas(p,:); species = species(p);

Train a discriminant analysis classifier by using measurements in the first half of the data.

half = floor(numObs/2); training = meas(1:half,:); trainingSpecies = species(1:half); Mdl = fitcdiscr(training,trainingSpecies);

Predict labels for the measurements in the second half of the data by using the trained classifier.

sample = meas(half+1:end,:); grouphat = predict(Mdl,sample);

Specify the group order and display the confusion matrix for the resulting classification.

group = species(half+1:end); [C,order] = confusionmat(group,grouphat,'Order',{'setosa','versicolor','virginica'})

`C = `*3×3*
29 0 0
0 22 2
0 0 22

`order = `*3x1 cell*
{'setosa' }
{'versicolor'}
{'virginica' }

The confusion matrix shows that the measurements belonging to setosa and virginica are classified correctly, while two of the measurements belonging to versicolor are misclassified as virginica. The output `order`

contains the order of the rows and columns of the confusion matrix in the sequence specified by the group order` {'setosa','versicolor','virginica'}`

.

### Confusion Matrix for Classification Using Tall Arrays

Perform classification on a tall array of the `fisheriris`

data set, compute a confusion matrix for the known and predicted tall labels by using the `confusionmat`

function, and plot the confusion matrix by using the `confusionchart`

function.

When you perform calculations on tall arrays, MATLAB® uses either a parallel pool (default if you have Parallel Computing Toolbox™) or the local MATLAB session. If you want to run the example using the local MATLAB session when you have Parallel Computing Toolbox, you can change the global execution environment by using the `mapreducer`

function.

Load Fisher's iris data set.

`load fisheriris`

Convert the in-memory arrays `meas`

and `species`

to tall arrays.

tx = tall(meas);

Starting parallel pool (parpool) using the 'local' profile ... Connected to the parallel pool (number of workers: 6).

ty = tall(species);

Find the number of observations in the tall array.

`numObs = gather(length(ty)); % gather collects tall array into memory`

Set the seeds of the random number generators using `rng`

and `tallrng`

for reproducibility, and randomly select training samples. The results can vary depending on the number of workers and the execution environment for the tall arrays. For details, see Control Where Your Code Runs.

rng('default') tallrng('default') numTrain = floor(numObs/2); [txTrain,trIdx] = datasample(tx,numTrain,'Replace',false); tyTrain = ty(trIdx);

Fit a decision tree classifier model on the training samples.

mdl = fitctree(txTrain,tyTrain);

Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 2: Completed in 3.9 sec - Pass 2 of 2: Completed in 1.5 sec Evaluation completed in 7.3 sec Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 4: Completed in 0.88 sec - Pass 2 of 4: Completed in 1.6 sec - Pass 3 of 4: Completed in 4 sec - Pass 4 of 4: Completed in 2.7 sec Evaluation completed in 11 sec Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 4: Completed in 0.54 sec - Pass 2 of 4: Completed in 1.2 sec - Pass 3 of 4: Completed in 3 sec - Pass 4 of 4: Completed in 2 sec Evaluation completed in 7.6 sec Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 4: Completed in 0.51 sec - Pass 2 of 4: Completed in 1.3 sec - Pass 3 of 4: Completed in 3.1 sec - Pass 4 of 4: Completed in 2.5 sec Evaluation completed in 8.5 sec Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 4: Completed in 0.42 sec - Pass 2 of 4: Completed in 1.2 sec - Pass 3 of 4: Completed in 3 sec - Pass 4 of 4: Completed in 2.1 sec Evaluation completed in 7.6 sec

Predict labels for the test samples by using the trained model.

txTest = tx(~trIdx,:); label = predict(mdl,txTest);

Compute the confusion matrix for the resulting classification.

tyTest = ty(~trIdx); [C,order] = confusionmat(tyTest,label)

C = M×N×... tall array ? ? ? ... ? ? ? ... ? ? ? ... : : : : : : Preview deferred. Learn more. order = M×N×... tall array ? ? ? ... ? ? ? ... ? ? ? ... : : : : : : Preview deferred. Learn more.

Use the `gather`

function to perform the deferred calculation and return the result of `confusionmat`

in memory.

gather(C)

Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 1: Completed in 1.9 sec Evaluation completed in 2.3 sec

`ans = `*3×3*
20 0 0
1 30 2
0 0 22

gather(order)

Evaluating tall expression using the Parallel Pool 'local': Evaluation completed in 0.032 sec

`ans = `*3×1 cell*
{'setosa' }
{'versicolor'}
{'virginica' }

The confusion matrix shows that three measurements in the versicolor class are misclassified. All the measurements belonging to setosa and virginica are classified correctly.

To compute and plot the confusion matrix, use `confusionchart`

instead.

cm = confusionchart(tyTest,label)

Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 1: Completed in 0.34 sec Evaluation completed in 0.6 sec Evaluating tall expression using the Parallel Pool 'local': - Pass 1 of 1: Completed in 0.48 sec Evaluation completed in 0.67 sec

cm = ConfusionMatrixChart with properties: NormalizedValues: [3×3 double] ClassLabels: {3×1 cell} Show all properties

## 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. 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. 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 and logical vectors, the order is the sorted order of

`s`

.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 and logical vectors, the order is the sorted order of

`s`

.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.

## Extended Capabilities

### Tall Arrays

Calculate with arrays that have more rows than fit in memory.

This function fully supports tall arrays. For more information, see Tall Arrays.

## Version History

**Introduced in R2008b**

## Open Example

You have a modified version of this example. Do you want to open this example with your edits?

## 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)