Two-Steps Benchmarks for Time Series Disaggregation
Generating a clone for bflSmooth_matrices_impl
Extrapolation function for the hfserie in a threeRuleSmooth
Smooth a time series
Default color palette
Default linetype palette
Default margins
Default ggplot theme
Virtual Class "disaggR" Class of disaggregations
Two-Steps Benchmarks for Time Series Disaggregation
Distance computation for disaggregations
Extend tsp with lf
Comparing a disaggregation with the high-frequency input
Comparing two disaggregations together
Producing the in sample predictions of a prais-lm regression
Comparing the inputs of a praislm regression
Extracting all the arguments submitted to generate an object
Extracting the standard error
Plotting disaggR objects
Extracting the regression of a twoStepsBenchmark
Producing a report
Extrapolation function for the residuals in a twoStepsBenchmark
Using an estimated benchmark model on another time series
A shiny app to reView and modify twoStepsBenchmarks
Extracting the autocorrelation parameter
Extracting the standard error
Extracting the smoothed part of a twoStepsBenchmark
Extracting the rate of a threeRuleSmooth
Bends a time series with a lower frequency one by smoothing their rate
Regress and bends a time series with a lower frequency one
The twoStepsBenchmark() and threeRuleSmooth() functions allow you to disaggregate a low-frequency time series with higher frequency time series, using the French National Accounts methodology. The aggregated sum of the resulting time series is strictly equal to the low-frequency time series within the benchmarking window. Typically, the low-frequency time series is an annual one, unknown for the last year, and the high frequency one is either quarterly or monthly. See "Methodology of quarterly national accounts", Insee Méthodes N°126, by Insee (2012, ISBN:978-2-11-068613-8, <https://www.insee.fr/en/information/2579410>).