TypeError: 'float' object cannot be interpreted as an index

Question

I am facing an issue which was not occurring before, maybe some rules have been changed.

    Traceback (most recent call last)
    <ipython-input-3-f47687e192a7> in <module>()
          5 n_examples = X.shape[0]
          6 n_train = n_examples * 0.5
    ----> 7 train_idx = np.random.choice(range(0,n_examples), size=n_train, replace=False)
          8 test_idx = list(set(range(0,n_examples))-set(train_idx))
          9 X_train = X[train_idx]

    mtrand.pyx in mtrand.RandomState.choice (numpy/random/mtrand/mtrand.c:18822)()

    TypeError: 'float' object cannot be interpreted as an index

Answer 1

The issue might be with the range function, which comes with Python. Its arguments must be integers. n_train gets turned to a float when n_examples is multiplied by 0.5. You just need to reconvert it to an int like int(n_examples * 0.5). This is actually the correct thing to do. If you'd have 11 examples, it wouldn't make sense to have 5.5 training and test examples.