This page allows performing nonlinear regressions (nonlinear least squares fittings). Because nonlinear optimization methods can be applied to any function, for the relation between two variables, it finds functions that best fit a given set of data points from a list of more than 100 functions, which include most common and interesting functions, like gaussians, sigmoidals, rationals, sinusoidals... Results are ordered then by the residual sum of squares and shown from the best suiting to the worst suiting case. Results can have a -usually small- deviation from optimality, but usually they are good and further improvement possibilities are small. NLR is a first step in the goal of automated scientific research.

Note:
If a lot of data points and a high number of parameters are selected, calculations can take some seconds.

• Copy & Paste: You can copy and paste data directly from a spreadsheet or a tabulated data file in the box below. Any character that cannot be part of a number -space, comma, tabulation...- is considered a column separator. By default commas are considered column separators; in the case you are using them as decimal separators check the option below. The exponent can be indicated by preceding it by the character E or e, as you can see in the example. Data must consist of two columns, x and y, to get the nonlinear regression y=f(x)._{ }

• Insert manually & See details: If you prefer you can insert all the points manually, for which you first have to specify the number of data points. You also can see details of the calculation -as the calculated value of y and the error of the best suiting function at each point- in this area.