search

robjhyndman / forecast

Forecasting Functions for Time Series and Linear Models
http://pkg.robjhyndman.com/forecast
1.12k stars 342 forks source link

No suitable ARIMA model found #716

Closed sakshisharma28 closed 6 years ago

sakshisharma28 commented 6 years ago

I have data on a daily basis. I created dummy variables for weekdays (mon, tue, wed, etc.) and holidays (just one column- 0,1)

My model is : modArima <- auto.arima(x, xreg = cbind(holiday_train, weekday_train))

It give me the following error msg: Error in auto.arima(x, xreg = cbind(holiday_train, weekday_train)) : No suitable ARIMA model found

colSums(weekday_train) Monday Tuesday Wednesday Thursday Friday Saturday Sunday 65 65 65 65 65 65 65

Where am I going wrong?

robjhyndman commented 6 years ago

You should have one fewer columns than days of the week. See https://otexts.org/fpp2/useful-predictors.html#seasonal-dummy-variables

sakshisharma28 commented 6 years ago

Thanks