Skip to main content

Python code to compute and plot (truncated, weighted) area under gain curves (agc)

Project description

AGC - Area under Gain Curves

Python code to compute and plot (truncated, weighted) area under gain curves (AGC)

For binary classification, gain curves are a nice alternative to ROC curves in that they can naturally be truncated to focus on the top scoring points only. Moreover, the data points can have weights. In this code, we provide three functions:

  • agc_score: Compute the area under the gain curve (AGC) for binary labelled data
  • gain_curve: Compute the proportion of data points and true positive rate for all thresholds, for plotting

The first function returns the normalized area by default (improvement over random, so this could be negative). The functions can be imported from the supplied agc.py file, or installed via pip install agc.

A simple example

## create toy binary labels and scores for illustration
labels = np.concatenate((np.repeat(1,100),np.repeat(0,900)))
scores = np.concatenate((np.random.uniform(.4,.8,100),np.random.uniform(.2,.6,900)))

## compute (normalized) area under the gain curve
print(agc_score(labels, scores))

## compute (un-normalized) area under the gain curve
print(agc_score(labels, scores, normalized=False))

## now the area for the top scoring 10% of the points
print(agc_score(labels, scores, truncate=0.1))

## or top scoring 100 points
print(agc_score(labels, scores, truncate=100))

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for agc, version 0.0.3
Filename, size File type Python version Upload date Hashes
Filename, size agc-0.0.3-py3-none-any.whl (4.6 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size agc-0.0.3.tar.gz (3.7 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page