2026/04/30 by Mustafa Mahmoud Aboulsaad · 1 voice
Computer Science · Physics and Astronomy · #cond-mat.mtrl-sci #cs.MS #physics.data-an
paper · pdf · doi:10.48550/arxiv.2605.00172
arxiv published 2026/04/30 · arxiv updated 2026/05/29
Reliable parameter extraction from experimental data is essential for quantitative analysis across spectroscopy, diffraction, photoluminescence, chromatography, microscopy, and time-resolved measurements. However, nonlinear fitting often remains difficult to reproduce, especially when complex models, correlated parameters, uncertain derived quantities, and user-dependent fitting choices are involved. We present FitED, a Python-based desktop application for nonlinear fitting of one-dimensional scientific data that combines an accessible graphical interface with a transparent and flexible numerical backend. FitED supports conventional peak profiles, including Gaussian, Lorentzian, Pseudo-Voigt, and exact area-normalized Voigt functions, as well as arbitrary user-defined analytical models for broader experimental applications. The software integrates local and global-search-assisted optimization strategies, automated model initialization, repeated stability testing, parameter-correlation analysis, and covariance-based propagation of uncertainty for derived quantities. By combining interactive usability with uncertainty-aware analysis and structured export of fitting results, FitED provides a practical platform for reproducible and interpretable fitting of experimental data. The software is intended to support both routine analysis and advanced model evaluation while preserving the parameter-level control required by experimental researchers.