Function to choose optimal daily seasonal forecast model

DayWModelPicker(data, Outcome, DateVar, H.Horizon = 14)

Arguments

data	A daily tsibble.
Outcome	A valid variable name for the Outcome to be modelled in in `data`.
DateVar	A valid variable name for the time index in `data`.
H.Horizon	An integer for the forecast horizon/test subset of `data`.

Value

A list containing:

Accuracy.Table: accuracy for the forecast horizon against the test sample.
test: the test set
train: the training set
Model.Fits: the model fits
Model.Forecasts: the forecasts
Min.Model: the minimum model by MAE
Min.Report: the minimum model report
Min.Res.Plot NW: the gg_tsdisplay for the minimum model
Min.Forecast.Plot NW: a plot of the minimum forecast by MAE

Examples

data(SNWSDPDX)
DayWModelPicker(SNWSDPDX, TX, date, H.Horizon=14)
#> Warning: Current temporal ordering may yield unexpected results.
#> ℹ Suggest to sort by ``, `date` first.
#> Joining, by = c("YR", "MO", "Day", "TX", "TN", "PR", "SN", "date")
#> Series: TX 
#> Model: LM w/ ARIMA(0,1,3) errors 
#> 
#> Coefficients:
#>           ma1      ma2      ma3  fourier(K = 1)C1_7  fourier(K = 1)S1_7
#>       -0.2211  -0.2975  -0.1758             -0.0092             -0.3660
#> s.e.   0.0298   0.0302   0.0294              0.3355              0.3353
#> 
#> sigma^2 estimated as 31.27:  log likelihood=-3389.31
#> AIC=6790.62   AICc=6790.7   BIC=6820.53
#> $test
#> # A tsibble: 14 x 8 [1D]
#>       YR    MO Day      TX    TN    PR    SN date      
#>    <int> <int> <chr> <dbl> <dbl> <dbl> <dbl> <date>    
#>  1  2019    12 18       43    32  0.18     0 2019-12-18
#>  2  2019    12 19       57    38  0.64     0 2019-12-19
#>  3  2019    12 20       59    50  0.4      0 2019-12-20
#>  4  2019    12 21       51    43  0.88     0 2019-12-21
#>  5  2019    12 22       46    42  0.17     0 2019-12-22
#>  6  2019    12 23       48    35  0.09     0 2019-12-23
#>  7  2019    12 24       40    32 NA        0 2019-12-24
#>  8  2019    12 25       44    31  0.02     0 2019-12-25
#>  9  2019    12 26       46    27 NA        0 2019-12-26
#> 10  2019    12 27       45    37  0.01     0 2019-12-27
#> 11  2019    12 28       48    40  0        0 2019-12-28
#> 12  2019    12 29       48    40 NA        0 2019-12-29
#> 13  2019    12 30       48    40  0        0 2019-12-30
#> 14  2019    12 31       51    40  0.22     0 2019-12-31
#> 
#> $train
#> # A tsibble: 1,081 x 8 [1D]
#>       YR    MO Day      TX    TN    PR    SN date      
#>    <int> <int> <chr> <dbl> <dbl> <dbl> <dbl> <date>    
#>  1  2017     1 1        40    32  0.05  NA   2017-01-01
#>  2  2017     1 2        35    29  0      0   2017-01-02
#>  3  2017     1 3        34    27  0      0   2017-01-03
#>  4  2017     1 4        33    27  0      0   2017-01-04
#>  5  2017     1 5        35    17  0      0   2017-01-05
#>  6  2017     1 6        34    17  0      0   2017-01-06
#>  7  2017     1 7        30    24  0.02   0.4 2017-01-07
#>  8  2017     1 8        34    28  0.53   0   2017-01-08
#>  9  2017     1 9        41    30  0.28   0   2017-01-09
#> 10  2017     1 10       38    31  0.65   6.5 2017-01-10
#> # … with 1,071 more rows
#> 
#> $Model.Fits
#> # A mable: 1 x 10
#>                       `K = 1`                     `K = 2`
#>                       <model>                     <model>
#> 1 <LM w/ ARIMA(0,1,3) errors> <LM w/ ARIMA(0,1,3) errors>
#> # … with 8 more variables: K = 3 <model>, ARIMA <model>, ETS <model>,
#> #   NNET1 <model>, NNET2 <model>, NNET3 <model>, prophet <model>,
#> #   Combo1 <model>
#> 
#> $Model.Forecasts
#> # A fable: 140 x 4 [1D]
#> # Key:     .model [10]
#>    .model date              TX .mean
#>    <chr>  <date>        <dist> <dbl>
#>  1 K = 1  2019-12-18 N(43, 31)  43.5
#>  2 K = 1  2019-12-19 N(45, 50)  44.9
#>  3 K = 1  2019-12-20 N(46, 57)  45.8
#>  4 K = 1  2019-12-21 N(46, 60)  45.7
#>  5 K = 1  2019-12-22 N(46, 63)  45.9
#>  6 K = 1  2019-12-23 N(46, 66)  46.3
#>  7 K = 1  2019-12-24 N(46, 69)  46.5
#>  8 K = 1  2019-12-25 N(46, 72)  46.4
#>  9 K = 1  2019-12-26 N(46, 75)  46.1
#> 10 K = 1  2019-12-27 N(46, 78)  45.8
#> # … with 130 more rows
#> 
#> $Accuracy.Table
#> # A tibble: 10 x 10
#>    .model  .type    ME  RMSE   MAE   MPE  MAPE  MASE RMSSE  ACF1
#>    <chr>   <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#>  1 ARIMA   Test   2.26  5.55  3.80  3.73  7.43   NaN   NaN 0.468
#>  2 Combo1  Test   3.12  5.95  4.15  5.53  8.04   NaN   NaN 0.446
#>  3 ETS     Test   3.98  6.45  4.64  7.33  8.97   NaN   NaN 0.423
#>  4 K = 1   Test   2.34  5.56  3.80  3.90  7.41   NaN   NaN 0.478
#>  5 K = 2   Test   2.26  5.54  3.82  3.74  7.47   NaN   NaN 0.475
#>  6 K = 3   Test   2.33  5.56  3.81  3.88  7.43   NaN   NaN 0.481
#>  7 NNET1   Test   3.92  6.45  4.46  7.28  8.49   NaN   NaN 0.346
#>  8 NNET2   Test   3.70  6.16  4.32  6.98  8.34   NaN   NaN 0.417
#>  9 NNET3   Test   3.54  6.06  4.50  6.51  8.74   NaN   NaN 0.385
#> 10 prophet Test   4.15  6.37  4.93  7.73  9.63   NaN   NaN 0.337
#> 
#> $Min.Model
#> # A tibble: 1 x 10
#>   .model .type    ME  RMSE   MAE   MPE  MAPE  MASE RMSSE  ACF1
#>   <chr>  <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 K = 1  Test   2.34  5.56  3.80  3.90  7.41   NaN   NaN 0.478
#> 
#> $Min.Report
#> # A mable: 1 x 1
#>                       `K = 1`
#>                       <model>
#> 1 <LM w/ ARIMA(0,1,3) errors>
#> 
#> $Min.Res.Plot
#>