Maybe I'm missing it here, but which ones are the test sets? Aren't we only provided with Train.csv?
I'm having the same question, I think we are only given training sets. Test sets have been removed already from the sample. I'm not sure how are we going to test model built.
Well, you can use the training data and split it into train and test set... The task here is to predict the latest 4/6 days left blank.
You will see that in the Train.csv there are missing values for the last 4/6 days (depending on which field). You will need to predict the soil humidities for these last 4/6 days. We refer to the last 4/6 days as the test set.
Hope this helps.