Learn R Programming

autoencoder (version 1.1)

Sparse Autoencoder for Automatic Learning of Representative Features from Unlabeled Data

Description

Implementation of the sparse autoencoder in R environment, following the notes of Andrew Ng (http://www.stanford.edu/class/archive/cs/cs294a/cs294a.1104/sparseAutoencoder.pdf). The features learned by the hidden layer of the autoencoder (through unsupervised learning of unlabeled data) can be used in constructing deep belief neural networks.

Copy Link

Version

Install

install.packages('autoencoder')

Monthly Downloads

43

Version

1.1

License

GPL-2

Maintainer

Yuriy Tyshetskiy

Last Published

July 2nd, 2015

Functions in autoencoder (1.1)

training_matrix_N=5e3_Ninput=100

An example training set of images for training sparse autoencoder
autoencoder-package

Implementation of sparse autoencoder for automatic learning of representative features from unlabeled data.
autoencode

Train a sparse autoencoder using unlabeled data
predict

Predict outputs of a sparse autoencoder
autoencoder_Ninput=100_Nhidden=100_rho=1e-2

A trained autoencoder example with 100 hidden units
visualize.hidden.units

Visualize features learned by a sparse autoencoder
autoencoder_Ninput=100_Nhidden=25_rho=1e-2

A trained autoencoder example with 25 hidden units