CoFESNARX: CoFESNARX

In demonejackson/CofesSNARX: Seasonal NARX via Keras

CoFESNARX

Description

CoFESNARX returns a compiled NARX model via Keras

Usage

CoFESNARX(x, y, dim, KL2 = 0.01, BL2 = 0.01, dr_o = 0, act = "relu")

Arguments

y, x	numeric vectors for the target and exegenous variables
dim	numberic vector for the dimension of the deisred netowrk
KL2	the L2 kernel_regularizerregularization factor for hidden layer kernal regularization
BL2	the L2 bias_regularizer regularization factor for hidden layer bias regularization
dr_o	drop out rate used between hidden layers
act	activation funciton used, if not specified 'relu' used

Details

This function takes target variable series, exogenous variables series, and the number of desired lags for each group respectively and then returns the requested lag series for each in a single data set. Ydelay/Xdelay are vectors that holds the consecutive delays required for the target variable/exogenous series (ex 1:3 or 4:5)

Value

returns the target combined with the original and lagged exegenous variables in one database.

demonejackson/CofesSNARX documentation built on April 26, 2022, 5:39 a.m.