Chapter 2: Issue with threshold.py about buildingmachinelearningsystemswithpython HOT 5 CLOSED

Thanks for the message. I can imagine this happening if fit_model is called with unexpected arguments, but should work if called correctly. For example, it is used in seeds_threshold.py and it works on my side.

Can you please post a full example that triggers the error for you?

from buildingmachinelearningsystemswithpython.

Thanks for pointing in the right direction. What I had failed to notice earlier was that after calling the function, the arrays, pred and labels were having different lengths which was yielding a single boolean value. In fact, this is the case that is happening in the code snippet contained in chapter.py at the below location:

Here, pred is a boolean array of length 100 whereas is_virginica is a boolean array of length 150 which is generating a single bool value. In my opinion, we should add one more line after building is_virginica as below to select non-setosa labels only as we are using only non-setosa features:

is_virginica = is_virginica[~is_setosa]

from buildingmachinelearningsystemswithpython.

Is this line 69 in the current (second edition) version of chapter.py? If so, it runs correctly on my side.

Note that is_virginica is defined as is_virginica = (labels == 'virginica') after labels = labels[~is_setosa].

If you are using the first edition, I do remember that there were a few of these manipulations which were (unfortunately) not in the book.

from buildingmachinelearningsystemswithpython.

Unfotunately, I am using the first edition of the book. However, I am now able to execute the code without any errors. Thanks for all your support.

from buildingmachinelearningsystemswithpython.

Great that you solved your problem.

The code related updates to the book are all in this repository, so if you run into any issues, check here or get in touch (we're always happy to help).

from buildingmachinelearningsystemswithpython.