If I do the seasonality test it gives me less than 0.04 in the P value, but it is not capable of forecasting and the seasonality is lost more if I add more years
An ARIMA(0,0,1) is not nonstationary. And forecasts are constant after one period out. And there is no seasonality estimated.
Maybe you want a different model?
I only want to forecast with the rain data I have. I have performed power functions but it does not change the result; If there is another library where I can forecast non-stationary data, it would help me
I'm not sure what you've done. The picture shows an ARIMA(0,0,1), which is stationary. A nonstationary model would have the second argument (typically) be a 1.
I am using auto.arima to study which is the best model, and from there I start to do it
myarima<- auto.arima(serie_lluvia,trace =TRUE,stepwise = FALSE, approximation = FALSE, max.d = 0)
Setting max.d=0 only allows for stationary series. Is that what you want?
what I want is a model where I can graph the forecast without giving me a linear value 
Well, you can eliminate the max.d=0, although my guess is that won't matter.
You can use arima rather than auto.arima and choose higher-order parameters for the first argument.
This topic was automatically closed 42 days after the last reply. New replies are no longer allowed.
If you have a query related to it or one of the replies, start a new topic and refer back with a link.