Function to estimate daily time series models

Function to estimate daily time series models

DaysModelFitter(data, Outcome)

Arguments

data

A tsibble.
Outcome

A valid variable name for the Outcome to be modelled in in `data`.

Value

A mable containing:

  • K=1,2,3: ARIMA models with fourier(K=1,2,3)

  • ARIMA: an ARIMA model

  • ETS: an ETS model

  • NNETAR(K=1,2,3): the model fits

  • prophet.Linear: a prophet model with linear trend

  • prophet.Logis: a prophet model with logistic trend

  • Combo1: the average of ETS and ARIMA

Examples

data(fb_returns)
DaysModelFitter(fb_returns, daily.returns)

#> Warning: 1 error encountered for K = 1
#> [1] K must be not be greater than period/2
#> Warning: 1 error encountered for K = 2
#> [1] K must be not be greater than period/2
#> Warning: 1 error encountered for K = 3
#> [1] K must be not be greater than period/2
#> Warning: 1 error encountered for NNET1
#> [1] K must be not be greater than period/2
#> Warning: 1 error encountered for NNET2
#> [1] K must be not be greater than period/2
#> Warning: 1 error encountered for NNET3
#> [1] K must be not be greater than period/2
#> Warning: 1 error encountered for prophet.Linear
#> [1] 'origin' must be supplied
#> Warning: 1 error encountered for prophet.Logis
#> [1] 'origin' must be supplied
#> # A mable: 1 x 11
#>        `K = 1`      `K = 2`      `K = 3`                  ARIMA          ETS
#>        <model>      <model>      <model>                <model>      <model>
#> 1 <NULL model> <NULL model> <NULL model> <ARIMA(0,0,0) w/ mean> <ETS(A,N,N)>
#> # … with 6 more variables: NNET1 <model>, NNET2 <model>, NNET3 <model>,
#> #   prophet.Linear <model>, prophet.Logis <model>, Combo1 <model>