BLNN_Update: Update a BLNN object with sampled weights and...

In BLNNdevs/BLNN: Bayesian Learning for Artificial Neural Networks

Description

Allows the user to update the network object after examining the samples. If the samples drawn passed the diagnostic checks then we need to update the network object with the average value of the accetped samples after warmup. If evidence was used in training the hyper-parameters will be updated as well.

Usage

BLNN_Update(NET, fit, index = NULL)

Arguments

fit	The fitted object returned by BLNN_train.
index	The starting index used to compute the average estimated value of the newtork weights. The default is warmup+1
Network	A BLNN network object

Value

The Network object with trained weights and updated hyper-parameters.

BLNNdevs/BLNN documentation built on Dec. 10, 2019, 3:31 a.m.