API Docs for: 0.0.1
Show:

SdA Class

Defined in: lib/SdA.js:19
Module: n42

Stacked autoencoder class

Constructor

SdA

()

Defined in lib/SdA.js:19

Item Index

Properties

Methods

predict

(
  • x
)

Defined in lib/SdA.js:154

Predict label with training data

Parameters:

  • x Matrix

    input data

pretrain

(
  • lr
  • corruptionLevel
  • epochs
)

Defined in lib/SdA.js:106

Training hidden layers with unsupervised learning

Parameters:

  • lr Float

    learning rate

  • corruptionLevel Float

    the standard deviation which is used by denoised autoencoder

  • epochs Int

    the number of times of running gradient decent

train

(
  • lr
  • epochs
)

Defined in lib/SdA.js:138

Training logistics regression algorithm which is on output layer

Parameters:

  • lr Float

    learning rate

  • epochs Int

    the number of times of running gradient decent

Properties

nLayers

Int

Defined in lib/SdA.js:57

number of hidden layers

sigmoidLayers

Array

Defined in lib/SdA.js:46

hidden layers which activations are sigmoid function

x

Matrix

Defined in lib/SdA.js:30

input data. This type is defined in sylvester library

y

Matrix

Defined in lib/SdA.js:38

label data. This type is defined in sylvester library