2
$\begingroup$

Are there any good statistical non-parametric forecasting methods besides machine learning methods like neural networks/decision trees etc. for time series analysis ?

If so, are there any R packages that implement them ?

Improve this question
$\endgroup$
3
  • 3
    $\begingroup$ How about non-parametric regression in the np package? $\endgroup$
    – chandler
    Commented Nov 22, 2013 at 17:31
  • 2
    $\begingroup$ There are exponential smoothing methods that can be used for forecasting. time series. They can be shown to be equivalent to time series models in the ARIMA framework so some would argue that they are therefore parametric. But one can also regard them as non-parametric if one uses them directly without estimating the equivalent underlying ARIMA model. There is simple ES, double ES and holt-winters ES. Each one makes different assumptions about the mean, the trend and the seasonality. $\endgroup$
    – mlofton
    Commented Aug 23, 2018 at 19:19
  • $\begingroup$ @chandler Can you elaborate on how to implement them ? $\endgroup$
    – noob
    Commented Sep 18, 2021 at 8:52

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.