Apply lambda function to numpy array elementwise doesnt work

Question

I am not quite sure why this isn´t working.. any explanation is highly appreciated. Mapping a function to np.array like:

test = np.array([0.6,0.7,1,0,0.5,0.2,0.4,0.3])
decision_boundary=0.6
decision_func = lambda x: 1. if (x >= decision_boundary) else 0.
decision_func(test)

results in the following value-error:

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

result = test>=decision_boundary will give you a boolean numpy array. result2=result.astype(np.int8) should switch it to 1s and 0s. — luthervespers, Feb 09 '21 at 00:09
I was rather curious why it doesnt work, but thanks for the suggestions anyway! — Bennimi, Feb 09 '21 at 00:12

score 1 · Accepted Answer · answered Feb 09 '21 at 00:06

You're not applying the function to each element of the array. You're applying the function to the whole array, and numpy is rightly telling you you can't convert a numpy array to Boolean to be used as the condition of if. Fortunately, in your case, >= is already vectorized on numpy arrays, so you can just do

x >= decision_boundary

Or if you really want ones and zeroes,

1 * (x >= decision_boundary)

In general, if you have a function that's not vectorized, you can make it so with the numpy function vectorize.

Apply lambda function to numpy array elementwise doesnt work

1 Answers1