I have metrological data they strictly have range -1 to 1. For example


measurement values are tied to a point in time .Example

Date    metr
9.27.2019 0:00:00   -0.97654
9.27.2019 1:00:00       0.0376544
9.27.2019 2:00:00       0.049765976
9.27.2019 3:00:00       -0.578
9.27.2019 4:00:00   0.045687
9.28.2019 5:00:00   -0.045688

I need to forecast the value by next 24 hours. Arima model doesn't work for me. because it is decimal fractions in a certain range from -1 to 1

What i should apply method to work to value with such specific. (ideally in R language)

  • 1
    $\begingroup$ Is it possible for values to be exactly -1 or +1? More details about the nature of the measurements would help, as would an example plot of values over time or a larger sample of data. From what you show the values seem to jump around a lot, making me wonder whether forecasting will be reliable. Might the underlying process be changing faster than the 1-hour spacings of your measurements? $\endgroup$ – EdM Sep 28 at 17:06
  • $\begingroup$ I’m not sure what you mean about ARIMA not working but you could multiply by 100 to avoid the decimal less than 1 issue. $\endgroup$ – Chris Umphlett Sep 29 at 2:18
  • $\begingroup$ See otexts.com/fpp2/limits.html $\endgroup$ – Rob Hyndman Sep 29 at 6:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.