in_convex_hull.py

import numpy as np
from scipy.spatial import Delaunay

training_set_labels = #.... from somewhere else

test_set_labels = np.atleast_2d(test_set_labels)

# construct convex hull of the training set using all labels
hull = Delaunay(training_set_labels)

# in_convex_hull will give true/false as to whether the test 
# set labels fall within the multidimensional convex hull of 
# the training set labels.
in_convex_hull = hull.find_simplex(test_set_labels) >= 0