TWC BioSearch
Discount Research Software Agent

Fit-Regression Analysis

The command macros on FIT.XLM perform a nonlinear regression procedure in order to obtain least squares estimates of up to 6 parameters in a nonlinear fit function Y=F(X,Par1.. Par6) for a set of (X,Y) data. The best fitting parameters are computed iteratively from a set of starting parameters using a Newton-Marquard algorithm.
After the best fit is found, the program calculates the (±) standard errors for the parameter estimates. The goodness of the fit is indicated by the average (Err%) deviation between observed and calculated data. This number is calculated as square root of the ratio of the averaged squared differences (observed - calculated) over the average of the squared observed data :

The macros on FIT allow plotting of the nonlinear regression model results (if desired, several on one chart) including error bars.
A trend function calculates the trend Y=F(X,Par1...Par6), or its inverse X=F-1(Y,Par1.. Par6) on a highlighted worksheet area, thus permitting the use of FIT for nonlinear calibration curves. FIT applies a root-finding Newton algorithm for the computation of the inverse function which eliminates the need to rearrange variables.
User-defined fit functions can be edited, recalled and stored for future use, together with user-assigned parameter names.

Requires : Windows 3.1, 95, 98 or NT and Excel 4.0 or higher.
 
P/N A11035 Single User
P/N A11035 5 User License 
P/N A11035 15 User License
P/N A21035 Macintosh version

Return to the Statistics, Math & Presentation section
 
 

TWC Biosearch International Home Page