At this moment, I have a nonlinear dataset that has X and Y-axis values. I want to come up with a nonlinear equation that represents the most values of this function without using a moving average filter.
It is a curve-fitting problem. Try using the Origin graphing software by OriginLab, or the MATLAB Curve-Fitting Toolbox by MathWorks. If you show the trend of the data, we can probably suggest some nonlinear models to fit the data.
use symbolic regression via Turing bot software. this software provides all of the equations regressing your data set in the space of mathematics with their corresponding errors and you can decide to choose the most accurate one