compute_error.py

def compute_error(partition):
    """Find the majority category of the partition. Count how big
       a part of the partition that the majority category is.
       Return the ratio of things that are *not* of the majority category."""
    categories = map(get_category, partition)
    majority = max(set(categories), key=categories.count)
    return 1 - (categories.count(majority) / float(len(partition)))

# This is what a typical partition looks like. The inner tuple may
# be very large (in the thousands) and the list of the length may
# be a couple of thousand.
test_partition = [((1, 2), 0), ((2, 3), 0), ((3, 2), 0), ((3, 5), 0),
                  ((4, 1), 1), ((4, 4), 1), ((5, 4), 1), ((6, 3), 1),
                  ((6, 5), 1), ((2, 7), 1)]

compute_error(test_partition)  # => 0.4