All functions

addFixedParsToLedger()

Insert the fixed hillSlope1 and inflectionPoint1 into the ledger's parameter sets

bestParamSets()

The parameter sets the ledger records for this ELF

calibratePowerLawFor()

Calibrate the "powerLaw" link on the ELF's flammable map

crossValidateSpread()

Two-fold cross-validation of the spread fit

crossValidateSpreadOneFold()

One cross-validation fold, run on its own (heldOutFold)

deriveCovMinMax()

Minimum and maximum of each covariate, for rescaling

doEvent.fireSense_spreadFit()

Event dispatcher for fireSense_spreadFit

.elfSeed()

A seed that depends only on the ELF

.inputObjects()

Default inputs

drawActivePars()

Draw parameter sets from the logistic's active range, by rejection

emulateObjective()

Search a noisy objective with a Gaussian-process emulator and return its best evaluated points

emulatorAtBoxEdge()

Whether a point lies in the trust region's widened margin

emulatorDefaults()

The emulator's settings and their defaults

emulatorNoiseSD()

The objective's noise near the best points

emulatorObjectiveArgs()

The objective's arguments, from the fit's runDEoptim() arguments

emulatorSpace()

The space the emulator searches: the full bounds, as a unit cube

emulatorSpreadPopulation()

The starting population: good evaluated points, spread out

emulatorStartPopulation()

Run the emulator on the fit's objective, on a cluster built as the fit's is

emulatorTrustBox()

The local phase's trust region

emulatorTrustStep()

One trust-region round's outcome

estimateSpreadParams()

Default upper or lower bounds for DEoptim

feasibleInitialPopulation()

A DEoptim starting population the objective can fit from

fitAndScoreFold()

Fit one cross-validation fold's complement and score its held-out years

fitSpread()

Fit the spread model

forkBudget()

How many forks this process may start now

fuelColumns()

Names of the fuel biomass columns among the non-annual covariates

hasPreRunFitForThisPolygon()

Does the shared ledger already hold a fit for the polygon this run fits?

histOfCovariates()

Histograms of the spread-fit covariates

inSampleSizes()

The in-sample fire-size validation of a fit

jobNeedsFitData()

Does this job read the data spreadFitPrepare makes?

ledgerWriteFile()

The ledger file a fit is written to

localSingleThreadBLAS()

One BLAS thread until the calling function returns

makeFitDiagnostics()

Turn the fit's re-score, profile and simulations into sim objects

mustFitThisPolygon()

Does this run fit the polygon, or use the ledger's fit? Stops when doFit is FALSE and there is no fit, and when it would fit a run over several ELFs.

needsPowerLawCalibration()

Does this job use the "powerLaw" calibration?

powerLawGammaOf()

The "powerLaw" link's gamma for this fit

profileLedgerFit()

Profile the ledger's fit of this ELF, without fitting

runSpreadWithoutDEoptim()

Evaluate the spread objective function without DEoptim

selectEmulatorBatch()

One batch of points to evaluate, chosen on the GP

spreadFitDEArgs()

The arguments fitSpread() gives fireSenseUtils::runDEoptim()

spreadFitLedgerRow()

The ledger row of a fit

spreadFitLoci()

Start cells and target sizes of the fires the spread model is fitted to

spreadFitPrep()

Prepare the inputs of the fit

spreadFitValidationFigures()

Figures comparing a spread fit with its data

spreadGates()

The spread objective's gates, as a function of parameter sets

startPopulation()

The population DEoptim starts from

thresholdForks()

How many forks the fit's start population and the "powerLaw" calibration may use

useEmulator()

Whether this fit starts from the emulator

withObjectiveCluster()

Run code on a cluster that holds the spread objective's data, built as the fit's is