What is the minimal size of input data

[chenx2018]

Hi,

I have 8 monthly time series in my input data set. When I called the function train_metalearning, I got the following error.

Auto calculated chunk size: 4 From 1 to 4 , 50 % of the dataset processed, remaining time: 5.64 seconds From 5 to 8 , 100 % of the dataset processed, remaining time: 0 seconds Calulating Average Naive2 errors for OWA... Error in hyperparameter_search(train_dataset, objective = objective, n_iter = 5, : Not enough data to do the crossvalidation! In addition: Warning message: The chosen seasonal unit root test encountered an error when testing for the second difference. From stl(): series is not periodic or has less than two periods 1 seasonal differences will be used. Consider using a different unit root test.

May I know what the minimal size for the input time series data to make train_metalearning() function work? Thanks!