Error

Module 7. Evaluating model performance Choice of cross-validation
#1

Hi,
I encountered this error when plotting the target_prediction. How could I solve this?
Thank you

image
image867×391 25 KB

#2

It does not make sense to plot the target_train and target_test in the same figure because they do not share the same x axis (as they do not have the same number of elements and do not represent the same dates).

Try plotting those in different notebook cells.

#3

Hi it works in the workbook titled ’ Non i.i.d. data’ though?
image

#4

Indeed I understand the source of confusion now. In the “Non i.i.d. data” notebook, target_train, target_test and target_predicted are three pandas Series with overlapping time spans as index for the x axis of the figure. This is probably not the case in your cell.

You can introspect the index of a pandas Series or DataFrame by typing:

target_train.index

target_test.index

target_predicted.index

in 3 different cells of your notebook to understand what’s going on.

#5

This question reveals a problem in the way we implicitly use indexes for plotting in the non-i.i.d. notebook.

We might want to reconsider this to avoid introducing complexity related to pandas indices on top of the complexity of the main message of this module which is a bout CV strategies.