predict

Class: Autoencoder

(To be removed) Reconstruct the inputs using trained autoencoder

predict 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

Y = predict(autoenc,X)

Description

Y = predict(autoenc,X) returns the predictions Y for the input data X, using the autoencoder autoenc. The result Y is a reconstruction of X.

Input Arguments

autoenc — Trained autoencoder

Autoencoder object

Trained autoencoder, specified as an Autoencoder object.

Xnew — Input data

matrix | cell array of image data | array of single image data

Input data, specified as a matrix of samples, a cell array of image data, or an array of single image data.

If the autoencoder autoenc was trained on a matrix, where each column represents a single sample, then Xnew must be a matrix, where each column represents a single sample.

If the autoencoder autoenc was trained on a cell array of images, then Xnew must either be a cell array of image data or an array of single image data.

Data Types: single | double | cell

Output Arguments

Y — Predictions for the input data Xnew

matrix | cell array of image data | array of single image data

Predictions for the input data Xnew, returned as a matrix or a cell array of image data.

• If Xnew is a matrix, then Y is also a matrix, where each column corresponds to a single sample (observation or example).

• If Xnew is a cell array of image data, then Y is also a cell array of image data, where each cell contains the data for a single image.

• If Xnew is an array of a single image data, then Y is also an array of a single image data.

Examples

Predict Continuous Measurements Using Trained Autoencoder

Load the training data.

X = iris_dataset;

The training data contains measurements on four attributes of iris flowers: Sepal length, sepal width, petal length, petal width.

Train an autoencoder on the training data using the positive saturating linear transfer function in the encoder and linear transfer function in the decoder.

autoenc = trainAutoencoder(X,'EncoderTransferFunction',...
'satlin','DecoderTransferFunction','purelin');

Reconstruct the measurements using the trained network, autoenc.

xReconstructed = predict(autoenc,X);

Plot the predicted measurement values along with the actual values in the training dataset.

for i = 1:4
h(i) = subplot(1,4,i);
plot(X(i,:),'r.');
hold on 
plot(xReconstructed(i,:),'go');
hold off;
end
title(h(1),{'Sepal';'Length'});
title(h(2),{'Sepal';'Width'});
title(h(3),{'Petal';'Length'});
title(h(4),{'Petal';'Width'});

The red dots represent the training data and the green circles represent the reconstructed data.

Version History

Introduced in R2015b

R2026a: predict will be removed

predict will be removed in a future release.

To learn more about autoencoder workflows, see these examples and topics:

• Build Networks with Deep Network Designer

• Time Series Anomaly Detection Using Deep Learning

• Detect Anomalies Using Wavelet Scattering with Autoencoders

• Train Variational Autoencoder (VAE) to Generate Images

• List of Deep Learning Layer Blocks and Subsystems

• Code Generation for Deep Learning Networks

For autoencoders specified as a dlnetwork object, make predictions using the predict and minibatchpredict functions.

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