GUM Forecast Learner

Generalised Univariate Model (GUM): a single-source-of-error state-space model with a user-defined transition matrix, persistence vector and measurement vector. Generalises exponential smoothing beyond the ETS structural template. Calls smooth::gum() from package smooth.

Dictionary

This mlr3::Learner can be instantiated via the dictionary mlr3::mlr_learners or with the associated sugar function mlr3::lrn():

mlr_learners$get("fcst.gum")
lrn("fcst.gum")

Meta Information

• Task type: “fcst”

• Predict Types: “response”, “quantiles”

• Feature Types: “logical”, “integer”, “numeric”

• Required Packages: mlr3, mlr3forecast, smooth

Parameters

Id	Type	Default	Levels
orders	untyped	c(1, 1)
lags	untyped	-
type	character	additive	additive, multiplicative
initial	character	backcasting	backcasting, optimal, two-stage, complete
persistence	untyped	NULL
transition	untyped	NULL
measurement	untyped	NULL
loss	character	likelihood	likelihood, MSE, MAE, HAM, MSEh, TMSE, GTMSE, MSCE, GPL
holdout	logical	FALSE	TRUE, FALSE
bounds	character	usual	usual, admissible, none
silent	logical	TRUE	TRUE, FALSE
regressors	character	use	use, select, adapt, integrate

References

Svetunkov I (2023). “Smooth forecasting with the smooth package in R.” 2301.01790, https://arxiv.org/abs/2301.01790.

See also

• Chapter in the mlr3book: https://mlr3book.mlr-org.com/chapters/chapter2/data_and_basic_modeling.html#sec-learners

• Package mlr3learners for a solid collection of essential learners.

• Package mlr3extralearners for more learners.

• Dictionary of Learners: mlr3::mlr_learners

• as.data.table(mlr_learners) for a table of available Learners in the running session (depending on the loaded packages).

• mlr3pipelines to combine learners with pre- and postprocessing steps.

• Package mlr3viz for some generic visualizations.

• Extension packages for additional task types:

  • mlr3proba for probabilistic supervised regression and survival analysis.

  • mlr3cluster for unsupervised clustering.

• mlr3tuning for tuning of hyperparameters, mlr3tuningspaces for established default tuning spaces.

Other Learner: LearnerFcst, mlr_learners_fcst.adam, mlr_learners_fcst.arfima, mlr_learners_fcst.arima, mlr_learners_fcst.auto_adam, mlr_learners_fcst.auto_arima, mlr_learners_fcst.auto_ces, mlr_learners_fcst.auto_gum, mlr_learners_fcst.auto_msarima, mlr_learners_fcst.auto_ssarima, mlr_learners_fcst.bagged, mlr_learners_fcst.bats, mlr_learners_fcst.ces, mlr_learners_fcst.croston, mlr_learners_fcst.elm, mlr_learners_fcst.es, mlr_learners_fcst.ets, mlr_learners_fcst.holt_winters, mlr_learners_fcst.mean, mlr_learners_fcst.mlp, mlr_learners_fcst.msarima, mlr_learners_fcst.nnetar, mlr_learners_fcst.prophet, mlr_learners_fcst.random_walk, mlr_learners_fcst.rlgt, mlr_learners_fcst.sma, mlr_learners_fcst.spline, mlr_learners_fcst.ssarima, mlr_learners_fcst.stlm, mlr_learners_fcst.struct_ts, mlr_learners_fcst.tbats, mlr_learners_fcst.theta, mlr_learners_fcst.tscount, mlr_learners_fcst.tslm

Super classes

mlr3::Learner -> mlr3::LearnerRegr -> LearnerFcst -> LearnerFcstSmooth -> LearnerFcstGum

Methods

Public methods

• LearnerFcstGum$new()

• LearnerFcstGum$clone()

Inherited methods

• mlr3::Learner$base_learner()
• mlr3::Learner$configure()
• mlr3::Learner$encapsulate()
• mlr3::Learner$format()
• mlr3::Learner$help()
• mlr3::Learner$predict()
• mlr3::Learner$predict_newdata()
• mlr3::Learner$print()
• mlr3::Learner$reset()
• mlr3::Learner$selected_features()
• mlr3::Learner$train()
• mlr3::LearnerRegr$predict_newdata_fast()

---

LearnerFcstGum$new()

Creates a new instance of this R6 class.

Usage

LearnerFcstGum$new()

---

LearnerFcstGum$clone()

The objects of this class are cloneable with this method.

Usage

LearnerFcstGum$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner and set parameter values
learner = lrn("fcst.gum")
print(learner)
#> 
#> ── <LearnerFcstGum> (fcst.gum): Generalised Univariate Model ───────────────────
#> • Model: -
#> • Parameters: list()
#> • Packages: mlr3, mlr3forecast, and smooth
#> • Predict Types: [response] and quantiles
#> • Feature Types: logical, integer, and numeric
#> • Encapsulation: none (fallback: -)
#> • Properties: exogenous, featureless, and missings
#> • Other settings: use_weights = 'error', predict_raw = 'FALSE'

# Define a Task
task = tsk("airpassengers")

# Create train and test set
ids = partition(task)

# Train the learner on the training ids
learner$train(task, row_ids = ids$train)

# Print the model
print(learner$model)
#> $model
#> Time elapsed: 0.04 seconds
#> Model estimated using fn() function: GUM(1[1],1[12])
#> With backcasting initialisation
#> Distribution assumed in the model: Normal
#> Loss function type: likelihood; Loss function value: 368.5502
#> Sample size: 96
#> Number of estimated parameters: 7
#> Number of degrees of freedom: 89
#> Information criteria:
#>      AIC     AICc      BIC     BICc 
#> 751.1004 752.3732 769.0509 771.9555 
#> 
#> $row_ids
#>  [1]  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
#> [26] 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
#> [51] 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
#> [76] 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96
#> 
#> $max_index
#> [1] "1956-12-01"
#> 

# Importance method
if ("importance" %in% learner$properties) print(learner$importance())

# Make predictions for the test rows
predictions = learner$predict(task, row_ids = ids$test)

# Score the predictions
predictions$score()
#> regr.mse 
#> 12441.94