karma.rfpermute: Autoregressive Random Forest with random permutations for...

In snarf-snarf/karma: Complementary tools for automated time-series machine learning.

Description

Autoregressive Random Forest with random permutations for variable selection

Usage

karma.rfpermute(y, ar_terms = NULL, seasonality = F, xreg = NULL,
  plot = T, stdout = F)

Arguments

ar_terms	Number of lagged terms to include as features
seasonality	Flag to indicate seasonal or nonseasonal series (T/F)
xreg	Features
plot	Show training convergence plot (T/F)
stdout	Print messages to standard output (T/F)
yt	Univariate time series object

Value

Trained forest with empirical p-values

See Also

tseries, forecast

Examples

yt = diff(log(JohnsonJohnson))

#Train model:
rf0 = tsml.rfPermute( y = yt, ar_terms = 15 )

#Get model p-values:
print(rf0$pval)

#Forecast:
tsml.forecast(rf0, h = 12, plot = T)

#Test 70-30:
tsml.cv(rf0, test_pct = 12, test_type = "window")

snarf-snarf/karma documentation built on May 24, 2019, 7:19 a.m.