Auto-ARIMA command: ValueError: y should be a 1d array, got an array of shape (90, 11) instead

Question

I'm currently trying to copy a code for my assignment:

stepwise_model = pm.auto_arima(df2, start_p=1, start_q=1,max_p=3, max_q=3, m=4, start_P=0, seasonal=True,
d=1, D=1, trace=True, error_action='ignore',suppress_warnings=True,stepwise=True)

In the assignment I have to use the above code, exept for the parameter m. I have changed the m =4, because 4 for quarterly data (which I need) In other words, use: start_p=1, start_q=1...

When I copied the code, changed into df2, I got an error:

Anyone can help me how I solve this error and run the auto-ARIMA command?

Answer 1

The error says that y (the input data) needs to be a 1 dimensional array since ARIMA only works on 1d arrays. It looks like the problem is that df2 is a 2 dimensional array, supposing it is a pandas dataframe with a time series you should only have the values of the series in it.