solverCan: A Mathematical Function Solver from Dataset Using Forward Differences, Derivative Ratio Analysis, and Iterative Residual Correction

This paper presents solverCan, a Python library that identifies mathematical functions from raw numerical data without requiring prior knowledge of the function type. The library employs a combination of forward difference analysis with Coefficient of Variation (CV) detection, Gaussian elimination for polynomial coefficient extraction, derivative ratio analysis for exponential function identification, and iterative residual correction for complex functions. Unlike machine learning approaches, solverCan produces explicit, human-readable equations and operates effectively with as few as 7–10 data points. The library automatically selects the best-fitting method among polynomial, exponential, iterative, and piecewise approaches, achieving less than 0.01% deviation for polynomial and exponential data, and less than 5% for trigonometric and irrational functions. The library is publicly available via PyPI (pip install solverCan).