Extreme Learning Machine Forecast Learner

Automatic time series forecasting with an extreme learning machine neural network, including automatic input lag selection, deterministic seasonality handling, and ensemble combination across multiple training repetitions. Calls nnfor::elm() from package nnfor.

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.elm")
lrn("fcst.elm")

Meta Information

• Task type: “fcst”

• Predict Types: “response”

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

• Required Packages: mlr3, mlr3forecast, nnfor, forecast

Parameters

Id	Type	Default	Levels	Range
m	integer	-		\([1, \infty)\)
hd	untyped	NULL		-
type	character	lasso	lasso, ridge, step, lm	-
reps	integer	20		\([1, \infty)\)
comb	character	median	median, mean, mode	-
lags	untyped	NULL		-
keep	untyped	NULL		-
difforder	untyped	NULL		-
sel.lag	logical	TRUE	TRUE, FALSE	-
direct	logical	FALSE	TRUE, FALSE	-
allow.det.season	logical	TRUE	TRUE, FALSE	-
det.type	character	auto	auto, bin, trg	-
barebone	logical	FALSE	TRUE, FALSE	-
retrain	logical	FALSE	TRUE, FALSE	-

References

Kourentzes, Nikolaos, Barrow, K. D, Crone, F. S (2014). “Neural network ensemble operators for time series forecasting.” Expert Systems with Applications, 41(9), 4235–4244. doi:10.1016/j.eswa.2013.12.011 .

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.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.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 -> LearnerFcstForecast -> LearnerFcstElm

Methods

Public methods

• LearnerFcstElm$new()

• LearnerFcstElm$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()

---

LearnerFcstElm$new()

Creates a new instance of this R6 class.

Usage

LearnerFcstElm$new()

---

LearnerFcstElm$clone()

The objects of this class are cloneable with this method.

Usage

LearnerFcstElm$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner and set parameter values
learner = lrn("fcst.elm")
print(learner)
#> 
#> ── <LearnerFcstElm> (fcst.elm): Extreme Learning Machine ───────────────────────
#> • Model: -
#> • Parameters: list()
#> • Packages: mlr3, mlr3forecast, nnfor, and forecast
#> • Predict Types: [response]
#> • Feature Types: logical, integer, numeric, character, factor, ordered,
#> POSIXct, and Date
#> • Encapsulation: none (fallback: -)
#> • Properties: featureless
#> • 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
#> ELM fit with 81 hidden nodes and 20 repetitions.
#> Series modelled in differences: D1.
#> Univariate lags: (3,4,7,8,10,12)
#> Deterministic seasonal dummies included.
#> Forecast combined using the median operator.
#> Output weight estimation using: lasso.
#> MSE: 76.9837.
#> 
#> $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 
#> 1613.816