fits: Fits the best model from classes ARIMA, ETS, TBATS and...

In IsmaelRohrig/Jurimetria: General Purpouse Tools for Jurimetrics

Description

Fits the best model from classes ARIMA, ETS, TBATS and NNETAR.

Usage

fits(x, train = 0.8, steps = NULL, max.points = 500,
  show.main.graph = T, show.sec.graph = F, show.value = T, PI = F,
  theme.doj = F)

Arguments

x	A vector or ts object.
train	The (initial) percentage of the time series to be used to train the models. Must be 0 < train < 1.
steps	Number of steps to forecast. If NULL, uses the number of points not used in training (testing points). Can't be less than the number of testing points.
max.points	Limits the maximum number of points to be used in modeling. Uses the first max.points points of the series.
show.main.graph	Logical. Should the main graphic (with the final model) be displayed?
show.sec.graph	Logical. Should the secondary graphics (with the training models) be displayed?
show.value	Logical. Should the values be displayed?
PI	Prediction Interval used in nnar models. May take long time processing.
theme.doj	Logical. Should the theme of Decades Of Jurimetrics be used?

Value

$fcast Predicted time series using the model that minimizes the forecasting mean square error.

$mse.pred Mean squared error of prediction. Used to decide the best model.

$runtime Running time.

References

Hyndman, R.J., & Athanasopoulos, G. (2018) Forecasting: principles and practice, 2nd edition, OTexts: Melbourne, Australia. otexts.com/fpp2.

https://robjhyndman.com/hyndsight/nnetar-prediction-intervals/

https://robjhyndman.com/talks/Google-Oct2015-part1.pdf

Zabala, F. J. and Silveira, F. F. (2019). Decades of Jurimetrics.

Examples

library(jurimetrics)

fits(livestock)
fits(livestock, theme.doj = T)
fits(livestock, show.main.graph = F, show.sec.graph = T, show.value = F)

fits(h02, .9)

fits(gas)

data('tjmg_year')
y1 <- ts(tjmg_year$count, start = c(2000,1), frequency = 1)
fits(y1)

data(tjrs_year_month)
y2 <- ts(tjrs_year_month$count, start = c(2000,1), frequency = 12)
fits(y2, train = 0.8, steps = 24)

IsmaelRohrig/Jurimetria documentation built on Nov. 23, 2019, 12:02 a.m.