patternnet

(To be removed) Generate pattern recognition network

patternnet will be removed in a future release. For more information, see Transition Legacy Neural Network Code to dlnetwork Workflows.

For advice on updating your code, see Version History.

Syntax

net = patternnet(hiddenSizes,trainFcn,performFcn)

Description

net = patternnet(hiddenSizes,trainFcn,performFcn) returns a pattern recognition neural network with a hidden layer size of hiddenSizes, a training function, specified by trainFcn, and a performance function, specified by performFcn.

Pattern recognition networks are feedforward networks that can be trained to classify inputs according to target classes. The target data for pattern recognition networks should consist of vectors of all zero values except for a 1 in element i, where i is the class they are to represent.

Examples

Construct and Train a Pattern Recognition Neural Network

This example shows how to design a pattern recognition network to classify iris flowers.

Load the training data.

[x,t] = iris_dataset;

Construct a pattern network with one hidden layer of size 10.

net = patternnet(10);

Train the network net using the training data.

net = train(net,x,t);

View the trained network.

view(net)

Estimate the targets using the trained network.

y = net(x);

Assess the performance of the trained network. The default performance function is mean squared error.

perf = perform(net,t,y)

perf = 
0.0181

classes = vec2ind(y);

Input Arguments

hiddenSizes — Size of the hidden layers

10 (default) | row vector

Size of the hidden layers in the network, specified as a row vector. The length of the vector determines the number of hidden layers in the network.

Example: For example, you can specify a network with 3 hidden layers, where the first hidden layer size is 10, the second is 8, and the third is 5 as follows: [10,8,5]

The input and output sizes are set to zero. The software adjusts the sizes of these during training according to the training data.

Data Types: single | double

trainFcn — Training function name

'trainscg' (default) | 'trainbr' | 'trainbfg' | 'trainrp' | 'trainlm' | ...

Training function name, specified as one of the following.

Training Function	Algorithm
'trainlm'	Levenberg-Marquardt
'trainbr'	Bayesian Regularization
'trainbfg'	BFGS Quasi-Newton
'trainrp'	Resilient Backpropagation
'trainscg'	Scaled Conjugate Gradient
'traincgb'	Conjugate Gradient with Powell/Beale Restarts
'traincgf'	Fletcher-Powell Conjugate Gradient
'traincgp'	Polak-Ribiére Conjugate Gradient
'trainoss'	One Step Secant
'traingdx'	Variable Learning Rate Gradient Descent
'traingdm'	Gradient Descent with Momentum
'traingd'	Gradient Descent

Example: For example, you can specify the variable learning rate gradient descent algorithm as the training algorithm as follows: 'traingdx'

Data Types: char

performFcn — Performance function

character vector

Performance function. The default value is 'crossentropy'.

This argument defines the function used to measure the network’s performance. The performance function is used to calculate network performance during training.

For a list of functions, in the MATLAB command window, type help nnperformance.

Output Arguments

net — Pattern recognition network

network object

Pattern recognition neural network, returned as a network object.

Version History

Introduced in R2010b

R2026a: patternnet will be removed

patternnet will be removed in a future release.

To train a neural network with tabular data, use one of these functions:

• fitrnet (Statistics and Machine Learning Toolbox) — Train regression neural network. Use this function for simple regression tasks.

• fitcnet (Statistics and Machine Learning Toolbox) — Train classification neural network. Use this function for simple classification tasks.

• trainnet — Train neural network. Use this function when the fitrnet and fitcnet functions do not provide the training options that you need for your task.

To train a neural network with time series data, use one of these options:

• Time Series Modeler app — Interactively design and train neural networks for time series data.

• trainnet — Train neural network. Use this function when you need more flexibility in the network architecture or training algorithm.

To design and customize your own neural network for these workflows, you can create a network using an array of deep learning layers or a dlnetwork object. To create and edit neural networks interactively and generate code, use the Deep Network Designer app. To learn more about neural network building workflows, see these examples and topics:

• Build Networks with Deep Network Designer

• List of Deep Learning Layers

• Example Deep Learning Network Architectures

• Multiple-Input and Multiple-Output Networks

• Define Custom Deep Learning Layers

For dlnetwork objects, to configure the hidden sizes of a pattern recognition network, adjust the outputSize argument of the fullyConnectedLayer objects. To specify the training algorithm, use the solverName argument of the trainingOptions function. To specify the loss function for training, use the lossFcn argument of the trainnet function.

For more information, see Transition Legacy Neural Network Code to dlnetwork Workflows.