Copatrec: A correlation pattern recognizer Python package for nonlinear relations

Abstract

Nonlinear regression plays a significant role as a method of analysis in different fields like statistics and simulations. While many researches focus on improving technical issues in the processes of nonlinear fitting, the absence of preprocessing, model selection, and validation is causing technical problems in various fields, i.e., statistical analysis and simulation methods. This work introduces a Python package named Copatrec, which facilitates preprocessing of data (i.e., data cleaning, standardization, and outlier detection), nonlinear model selection (from a group of complex behaviors’ mathematical expressions), and validation (by returning a result object that contains validation metrics and visualization functions), as observed in diverse scientific fields, including human-related sciences (e.g., social, political, economics, psychology, and health), and engineering fields like computer science (e.g., computer vision), electrical engineering, as well as fundamental science like physics and chemistry, where a precise characterization of the nonlinear relationship is central to the analytical outcome.

Keywords

• Complexity
• Nonlinear regression
• Correlation pattern recognition
• Python