Structural Time Series Forecast Learner

Structural time series model fit by maximum likelihood. Three model types are supported: local level, local linear trend, and basic structural model (level + trend + seasonal). Calls stats::StructTS() from package stats.

type = "BSM" requires a seasonal time series (frequency > 1). Prediction is performed via forecast::forecast.StructTS() which yields point forecasts and predictive intervals from the Kalman filter.

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.struct_ts")
lrn("fcst.struct_ts")

Meta Information

• Task type: “fcst”

• Predict Types: “response”, “quantiles”

• Feature Types: “logical”, “integer”, “numeric”, “character”, “factor”, “ordered”, “POSIXct”, “Date”

• Required Packages: mlr3, mlr3forecast, forecast

Parameters

Id	Type	Default	Levels
type	character	level	level, trend, BSM
init	untyped	NULL
fixed	untyped	NULL
optim.control	untyped	NULL
lambda	untyped	NULL
biasadj	logical	FALSE	TRUE, FALSE

References

Harvey AC (1989). Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge University Press, Cambridge.

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.gum, 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.tbats, mlr_learners_fcst.theta, mlr_learners_fcst.tscount, mlr_learners_fcst.tslm

Super classes

mlr3::Learner -> mlr3::LearnerRegr -> LearnerFcst -> LearnerFcstForecast -> LearnerFcstStructTS

Methods

Public methods

• LearnerFcstStructTS$new()

• LearnerFcstStructTS$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()

---

LearnerFcstStructTS$new()

Creates a new instance of this R6 class.

Usage

LearnerFcstStructTS$new()

---

LearnerFcstStructTS$clone()

The objects of this class are cloneable with this method.

Usage

LearnerFcstStructTS$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner and set parameter values
learner = lrn("fcst.struct_ts")
print(learner)
#> 
#> ── <LearnerFcstStructTS> (fcst.struct_ts): Structural Time Series ──────────────────────────────────────────────────────────────────────────────────────────────────────────────────
#> • Model: -
#> • Parameters: list()
#> • Packages: mlr3, mlr3forecast, and forecast
#> • Predict Types: [response] and quantiles
#> • Feature Types: logical, integer, numeric, character, factor, ordered, POSIXct, and Date
#> • Encapsulation: none (fallback: -)
#> • Properties: 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
#> 
#> Call:
#> StructTS(x = passengers)
#> 
#> Variances:
#>   level    slope     seas  epsilon  
#>    0.00    98.02    18.92     0.00  
#> 
#> $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 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 52 53 54 55 56 57 58
#> [59] 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 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 
#> 279831.3