Confidence Threshold Evaluation Comparison

confidence_roc.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""Uses multiple confidence thresholds to evaluate a model.

"""

##############################################################################
# Copyright (c) 2015 BigML, Inc
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
# LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
# WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
##############################################################################

import sys
import argparse

from bigml.api import BigML


def main(args=sys.argv[1:]):
    """Parses command-line parameters and calls the actual main function.

    """
    parser = argparse.ArgumentParser(
        description="Generates a ROC curve using confidence as threshold",
        epilog="BigML, Inc")

    # model
    parser.add_argument('--model',
                        required=True,
                        action='store',
                        dest='model',
                        help="Classification model")

    # test set
    parser.add_argument('--test',
                        required=True,
                        action='store',
                        dest='test',
                        help="Test set")

    # positive class
    parser.add_argument('--positive',
                        required=True,
                        action='store',
                        dest='positive',
                        help="Positive class")

    # negative class
    parser.add_argument('--negative',
                        required=True,
                        action='store',
                        dest='negative',
                        help="Negative class")

    args = parser.parse_args(args)

    api = BigML()

    thresholds = [confidence/10.0 for confidence in range(11)]

    evaluations = []

    for threshold in thresholds:
        evaluation = api.create_evaluation(args.model, args.test, {
            'name': "%s : %s" % (args.positive, threshold),
            'confidence_threshold': threshold,
            'positive_class': args.positive,
            'negative_class': args.negative})
        api.ok(evaluation)
        evaluations.append(evaluation['resource'])
        evaluation = api.create_evaluation(args.model, args.test, {
            'name': "%s : %s" % (args.negative, threshold),
            'confidence_threshold': threshold,
            'positive_class': args.negative,
            'negative_class': args.positive})
        api.ok(evaluation)
        evaluations.append(evaluation['resource'])

    url_param = [evaluation[len('evaluation/'):] for evaluation in evaluations]

    url_pattern = "labs/evaluationcomp/?evaluations=%s;positive_class=%s"
    url = url_pattern % (",".join(url_param), args.positive)

    # Use your VPC or private deployment URL here
    print "https://bigml.com/" + url

if __name__ == "__main__":
    main()