Does input X in degrade_dataset(...) has nan (missing values)?

[yezhengli-Mr9]

Hi Indro,
Does input X in degrade_dataset(...) possibly have nan (missing values, I mean the raw data has intrinsic missing values)?

If so, why mask_1d = np.ones(n) rather than mask_1d = #where X is not nan?

[spindro]

Hi!
If you use degrade_dataset it is supposed that the raw data is perfect and you want to create missing values.
By the way, if your raw data has already some missing values, you can still use degrade_dataset to add some noise. In this case, you have to sum the original mask of missing values to the one produced by the function

If so, why 'mask_1d = np.ones(n)' rather than 'mask_1d = #where X is not nan'?

What I said before should clear also this point.

Let me know if it is ok now :)

[Coding511]

@spindro Is this code working for filling missing values?