Metadata-Version: 2.1
Name: slope_er
Version: 0.1.4
Summary: Python implementation of Slope Error Rate (SER) by Arsalan Rahman Mirza.
Author-email: Arsalan Rahman Mirza <arsalan.mirza@soran.edu.iq>
Project-URL: Homepage, https://pypi.org/project/slope_er/
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE

Slope Error Rate (SER)

A Python implementation of Slope Error Rate (SER) a novel performance evaluation metric for binary classification, 
designed to provide improved sensitivity and robustness under class imbalance, particularly for biometric authentication systems.

Overview

Slope Error Rate (SER) is a distance-based evaluation metric that quantifies the trade-off between False Acceptance Rate (FAR) 
and False Rejection Rate (FRR) across decision thresholds. Unlike traditional metrics such as Equal Error Rate (EER), which rely 
on a single operating point, SER captures error dynamics and provides a more informative assessment of classifier behavior.

SER is especially useful in applications where:

Data is highly imbalanced
False positives and false negatives have asymmetric costs
Threshold sensitivity matters

A lower SER value indicates better overall performance.

Key Features
Designed for binary classification
Robust under class imbalance
Threshold-sensitive evaluation
Based on FAR and FRR geometry
Lightweight and easy to integrate
Suitable for biometrics, spoof detection, anomaly detection, and security systems

to install write

pip install slope-er==0.1.3

to use write

import slope_er


# Call the function
result = slope_er.ser(0.3, 0.7,True)

print(f"The Slope Error Rate is: {result}")


