auto.karma: Train ARIMA model for automated univariate time series...

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

Description

Improved ARIMA model selection using smart training algorithms and heuristic search options.

Usage

auto.karma(y, method = "best-first", xreg = NULL, metric = "AICc",
  cv = "out", max_iter = 5, test_pct = "auto", test_type = "auto",
  max_sdiff = F, violate_aic = T, karma_transform = F, plot = F,
  stdout = F)

Arguments

y	A univariate time-series vector; type <numeric> or <ts>
method	Model selection algorithm; "markov-selection": Trains MS-ARIMA model; "local-descent": Trains LD-ARIMA model; "random-walk": Trains RW-ARIMA model
xreg	Optional vector or matrix of exogenous regressors; see documentation for Arima(), package 'forecast'
metric	Choose a model validation metric that will be used as the main optimisation criterion during model selection
cv	Choose cross-validation dataset to be used during model selection; "out": Performance of out-of-sample forecast (classic train/test split) will be used for model validation; "in": Performance of in-sample forecast (classic parametric regression type of validation) will be used for model validation
max_iter	For karma.boxjenkins(): Maximum number of iterations without improvement before the algorithm converges naturally (reached a global or local optimum)
test_pct	Percentage of train-test split in cross-validation (e.g. 70-30), positive integer for "window" or "percentage" test_type; "auto" to read from karma.fit object or generate; negative integer value to set window size to a multiple of the series' frequency
test_type	Train-test split type, i.e. percentage or fixed window; "auto": will try to read from karma.fit object or generate; "percentage": test_pct = 12 will be read as the 12 percent of the length of the series; "window": test_pct = 12 will be read as the 12 last time points (e.g. months) of the series; "auto" if input series is a ts() object, test_type is set to "window" and test_pct is set to twice the frequency of the series - if test_pct is given a negative factor, then test_pct (window size) will be set to the frequency of the series times the absolute value of that negative number
max_sdiff	Option to set a threshold to the order of seasonal differencing. (T/F)
violate_aic	Heuristic option to escape local optima by preferring a model of lower AICc even if it's not comparable in theory; (T/F)
karma_transform	Option to automatically detect and apply the right transformation for wide-sense stationarity (T/F)
plot	Option to depict plots during local search; if TRUE (default), AC and PAC plots are active <logical>
stdout	Option to output optimisation diagnostics during local search; <logical>

Value

Object of class "karma.fit"; (extends class "Arima" from package 'forecast')

See Also

tseries, forecast

Examples

# One-line MS-ARIMA fit:
sfit = auto.karma(mdeaths)
# Validate trained model on test set:
karma.cv(sfit)
# One-line LD-ARIMA fit with in-sample validation and verbose search output:
sfit = auto.karma(y = mdeaths, method = "local-descent", cv = "in", stdout = T )
karma.cv(sfit)
# One-line RW-ARIMA using nonparametric training:
sfit = auto.karma(y = mdeaths, method = "random-walk", stdout = T )
karma.cv(sfit)

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