SeqMetrics: a unified library for performance metrics calculation in Python
Fazila Rubab, Sara Iftikhar, Ather Abbas · The Journal of Open Source Software · 2024
Current Python infrastructure lacks a robust, unified, and simplified library for error and performance metrics calculations.The SeqMetrics application responds to this critical need, providing a robust toolkit for evaluating regression and classification models with a comprehensive suite of performance metrics suitable for tabular and time series data.Designed for versatility, the library offers 112 regression and 22 classification metrics through both functional and class-based APIs.The design of the library ensures seamless integration into various coding environments.The web-based graphical user interface (GUI) of SeqMetrics enhances user accessibility, allowing efficient input through data arrays or file imports.It serves to users of varied programming expertise, offering a user-friendly interface for rigorous model assessment.The library prioritizes computational efficiency and has minimal dependencies with straightforward pip installation from PyPI.Rigorous testing of Seqmetrics ensures robustness, supported by extensive documentation for effective utilization by people from diverse backgrounds.Overall, SeqMetrics bridges the gap in Python's scientific analysis toolkit, contributing to data analysis.