Skip to contents

Temporal disaggregation of a time series by regression models.

Source: R/tempdisagg.R
temporal_disaggregation.Rd

Perform temporal disaggregation of low frequency to high frequency time series by regression models. Models included are Chow-Lin, Fernandez, Litterman and some variants of those algorithms.

Usage

temporal_disaggregation(
  series,
  constant = TRUE,
  trend = FALSE,
  indicators = NULL,
  model = c("Ar1", "Rw", "RwAr1"),
  freq = 4L,
  average = FALSE,
  rho = 0,
  rho.fixed = FALSE,
  rho.truncated = 0,
  zeroinitialization = FALSE,
  diffuse.algorithm = c("SqrtDiffuse", "Diffuse", "Augmented"),
  diffuse.regressors = FALSE,
  nbcsts = 0L,
  nfcsts = 0L
)

Arguments

series

The low frequency time series that will be disaggregated. It must be a ts object.

constant

Constant term (T/F). Only used with "Ar1" model when zeroinitialization = F.

trend

Linear trend (T/F, F by default)

indicators

High-frequency indicator(s) used in the temporal disaggregation. It must be a (list of) ts object(s).

model

Model of the error term (at the disaggregated level). "Ar1" = Chow-Lin, "Rw" = Fernandez, "RwAr1" = Litterman.

freq

Integer. Annual frequency of the disaggregated series. Ignored when an indicator is provided.

average

Average conversion (T/F). Default is F, which means additive conversion.

rho

(Initial) value of the parameter. Only used with Ar1/RwAr1 models.

rho.fixed

Fixed rho (T/F, F by default)

rho.truncated

Range for rho evaluation (in [rho.truncated, 1[)

zeroinitialization

The initial values of an auto-regressive model are fixed to 0 (T/F, F by default)

diffuse.algorithm

Algorithm used for diffuse initialization. "SqrtDiffuse" by default.

diffuse.regressors

Indicates if the coefficients of the regression model are diffuse (T) or fixed unknown (F, default)

nbcsts

Number of backcast periods. Ignored when an indicator is provided.

nfcsts

Number of forecast periods. Ignored when an indicator is provided.

Value

An object of class "JD3TempDisagg"

See also

temporal_interpolation,

temporal_disaggregation_raw for temporal disaggregation of atypical frequency series,

temporal_interpolation_raw for interpolation of atypical frequency series

Examples

# chow-lin with monthly indicator
Y <- rjd3toolkit::aggregate(rjd3toolkit::Retail$RetailSalesTotal, 1)
x <- rjd3toolkit::Retail$FoodAndBeverageStores
td <- temporal_disaggregation(Y, indicators = x)
td$estimation$disagg
#>           Jan      Feb      Mar      Apr      May      Jun      Jul      Aug
#> 1992 137916.2 127636.4 138517.7 143927.8 158328.5 149506.4 165994.4 157106.1
#> 1993 146290.9 128256.0 152066.7 156486.6 167812.5 163176.0 180412.0 163065.1
#> 1994 155985.0 139247.5 173398.9 166717.7 177111.7 180704.9 188521.6 182227.7
#> 1995 168520.9 149989.8 182344.2 177053.4 191096.4 190264.4 195173.8 193101.2
#> 1996 179435.3 170227.9 193110.6 184407.0 206647.8 197638.4 207403.2 211601.9
#> 1997 195237.9 169237.7 207992.6 190264.2 219075.2 200135.7 218598.0 217058.6
#> 1998 202353.7 174820.8 202058.5 207469.2 225420.4 211913.4 232254.9 222452.2
#> 1999 212635.3 191944.3 225220.7 221938.2 241654.4 229761.2 252098.9 234332.2
#> 2000 220459.3 214640.6 244542.4 243574.2 256653.5 254895.8 261298.1 256676.4
#> 2001 234869.1 217233.2 253501.6 241471.0 267666.8 259748.6 261438.3 266512.4
#> 2002 248121.9 224394.8 267193.4 239316.0 276996.6 260785.1 271586.4 273492.2
#> 2003 261953.1 232128.8 263355.1 261124.9 286605.5 266429.9 287538.3 283834.3
#> 2004 279072.3 251916.1 276322.0 279278.5 300072.3 285913.0 306671.3 288229.1
#> 2005 289075.7 261988.0 305037.9 292996.1 313415.5 308576.3 321094.4 312769.7
#> 2006 297912.7 279466.4 315389.8 311076.6 334737.8 326677.8 334816.9 333352.5
#> 2007 318771.0 293740.0 334611.7 316630.6 349743.6 340016.0 341018.3 341834.3
#> 2008 325379.5 306804.6 334155.8 315345.2 355478.9 328328.9 346321.1 341955.0
#> 2009 307285.7 260043.2 288734.1 293851.8 320224.1 297357.7 315867.3 305260.0
#> 2010 302650.2 276723.8 316851.2 308077.3 335229.0 318399.9 338574.3 326435.3
#>           Sep      Oct      Nov      Dec
#> 1992 144594.1 157550.7 148144.5 186493.2
#> 1993 157732.5 164226.6 160607.4 202115.8
#> 1994 175159.2 175396.9 176480.4 219069.5
#> 1995 182535.4 179420.3 186502.3 226501.9
#> 1996 185666.8 198697.8 203890.8 227937.5
#> 1997 196337.4 210469.8 208483.0 241112.8
#> 1998 209207.0 223073.5 214748.3 261333.0
#> 1999 231533.7 236719.8 234519.8 296197.4
#> 2000 245772.2 243067.1 251544.1 295632.2
#> 2001 248131.0 253908.4 261993.4 301251.1
#> 2002 245738.3 259914.1 270627.9 296155.3
#> 2003 261679.5 276245.4 276055.5 311203.6
#> 2004 285076.1 293043.8 292977.9 341857.6
#> 2005 304841.0 309399.8 312547.0 364949.5
#> 2006 316722.9 321745.1 330795.8 377441.7
#> 2007 321570.6 328295.3 339336.6 380229.9
#> 2008 309490.3 323626.9 320005.4 346041.4
#> 2009 289382.4 306533.4 302747.1 351184.0
#> 2010 318513.9 331406.2 334459.6 382144.2

# fernandez with/without quarterly indicator
data("qna_data")
Y <- ts(qna_data$B1G_Y_data[,"B1G_FF"], frequency = 1, start = c(2009,1))
x <- ts(qna_data$TURN_Q_data[,"TURN_INDEX_FF"], frequency = 4, start = c(2009,1))
td1 <- temporal_disaggregation(Y, indicators = x, model = "Rw")
td1$estimation$disagg
#>          Qtr1     Qtr2     Qtr3     Qtr4
#> 2009 4054.319 4546.626 4230.812 4722.644
#> 2010 3965.439 4546.603 4200.157 5007.901
#> 2011 4142.939 4873.447 4470.202 5369.411
#> 2012 4364.657 4907.514 4519.913 5057.616
#> 2013 4338.360 4796.134 4474.717 5104.289
#> 2014 4485.847 4841.382 4510.477 5170.294
#> 2015 4415.320 4996.023 4612.936 5269.721
#> 2016 4513.425 5095.867 4714.789 5354.319
#> 2017 4692.663 5195.768 4717.649 5543.821
#> 2018 4925.526 5552.812 5207.136 6081.925
#> 2019 5249.821 5784.822 5326.067 6153.489
#> 2020 5228.107 5044.753 5176.463 6117.177
#> 2021 5355.338 6047.411 5450.219 6546.932

td2 <- temporal_disaggregation(Y, model = "Rw", nfcsts = 6)
td2$estimation$disagg
#>          Qtr1     Qtr2     Qtr3     Qtr4
#> 2009 4394.912 4392.387 4387.338 4379.763
#> 2010 4369.663 4389.768 4440.077 4520.592
#> 2011 4631.312 4708.675 4752.682 4763.331
#> 2012 4740.623 4720.261 4702.244 4686.572
#> 2013 4673.246 4669.967 4676.736 4693.552
#> 2014 4720.416 4743.795 4763.690 4780.100
#> 2015 4793.025 4810.385 4832.180 4858.410
#> 2016 4889.074 4913.550 4931.838 4943.938
#> 2017 4949.849 4987.264 5056.182 5156.604
#> 2018 5288.529 5402.627 5498.899 5577.345
#> 2019 5637.965 5658.447 5638.791 5578.997
#> 2020 5479.065 5404.116 5354.151 5329.168
#> 2021 5329.168 5329.168 5329.168 5329.168
#> 2022 5329.168 5329.168                  

# chow-lin on index series
Y_index <- 100 * Y / Y[1]
x_index <- 100 * x / x[1]
td3 <- temporal_disaggregation(Y, indicators = x, average = TRUE)
td3$estimation$disagg
#>          Qtr1     Qtr2     Qtr3     Qtr4
#> 2009 16328.55 18098.49 17036.42 18754.14
#> 2010 16074.89 18100.48 16925.50 19779.53
#> 2011 16842.84 19414.67 18022.13 21144.36
#> 2012 17640.81 19526.21 18182.77 20049.00
#> 2013 17508.95 19107.93 18013.75 20223.37
#> 2014 18079.35 19318.66 18167.70 20466.30
#> 2015 17874.01 19895.30 18559.96 20846.74
#> 2016 18234.19 20279.41 18977.63 21222.38
#> 2017 18931.73 20688.01 19034.39 21945.48
#> 2018 19929.86 22152.34 20964.98 24022.42
#> 2019 21151.75 23057.83 21500.42 24346.80
#> 2020 20922.72 20273.19 20818.86 24251.23
#> 2021 21872.07 24423.57 22425.88 26281.08