I am trying to use the sktime python library for forecasting some time series data.
My x_train is of shape (277, 10) with the DateTime as index. My x_test is of shape (70, 10)
My code for using ARIMA is:
fh = ForecastingHorizon(y_test.index, is_relative=False)
arima = ARIMA()
arima.fit(y=y_train, X=x_train)
y_pred = arima.predict(fh=fh, X=x_test)
But the predict step is giving me error:
X array dims (n_rows) != n_periods. Received n_rows=70 and n_periods=80
I am totally unable to understand where the 80 is coming from. I couldn't find any explanation in the sktime documentation. The index of fh and x_test are matching.
I am also getting similar error in case of VARMAX, in case of which it is expecting (89,10) X. where are these numbers coming from? Why can't I predict for any period of time? What am I doing wrong here?