Curve Fitting Toolbox™

fitoptions - Create or modify fit options structure

Syntax

options = fitoptions
options = fitoptions(model)
options = fitoptions(model,fld1,val1,fld2,val2,...)
options = fitoptions('Method',method)
options = fitoptions('Method',method,fld1,val1,fld2,val2,...)
newoptions = fitoptions(options,fld1,val1,fld2,val2,...)
newoptions = fitoptions(options1,options2)

Description

options = fitoptions creates the default fit options structure options. Fields in the options structure, listed in the table below with their default values, are supported by all fitting methods.

options = fitoptions(model) creates the default fit options structure for the library or custom model specified by the string model. You can display library model names with the cflibhelp function.

options = fitoptions(model,fld1,val1,fld2,val2,...) creates a fit options structure for the specified model with the fields specified by the strings fld1, fld2, ... set to the values val1, val2, ..., respectively.

options = fitoptions('Method',method) creates the default fit options structure for the fitting method specified by the string method. Supported fitting methods are listed in the table below.

options = fitoptions('Method',method,fld1,val1,fld2,val2,...) creates the default fit options structure for the fitting method specified by the string method with the fields specified by the strings fld1, fld2, ... set to the values val1, val2, ..., respectively.

newoptions = fitoptions(options,fld1,val1,fld2,val2,...) modifies the existing fit options structure options by setting the fields specified by the strings fld1, fld2, ... set to the values val1, val2, ..., respectively. The new options structure is returned in newoptions.

newoptions = fitoptions(options1,options2) combines the input fit options structures options1 and options2 to create the output fit options structure newoptions. If the input structures have Method fields set to the same value, the nonempty values for the fields in options2 override the corresponding values in options1 in the output structure. If the input structures have Method fields set to different values, the output structure will have the same Method as options1, and only the values of the Normalize, Exclude, and Weights fields of options2 will override the corresponding values in options1.

Remarks

Field values in a fit options structure can be referenced with the get method and assigned with the set method. For example:

options = fitoptions('fourier1');
get(options,'Method')
ans =
NonlinearLeastSquares
get(options,'MaxIter')
ans =
   400
set(options,'Maxiter',1e3);
get(options,'MaxIter')
ans =
        1000

Field values can also be referenced and assigned using the dot notation. For example:

options.MaxIter
ans =
        1000
options.MaxIter = 500;
options.MaxIter
ans =
   500

Additional Fit Options

Additional fields in the fit options structure, beyond the default fields Normalize, Exclude, Weights, and Method, are available according to the fitting method.

If the Method field has the value 'NearestInterpolant', 'LinearInterpolant', 'PchipInterpolant', or 'CubicSplineInterpolant', there are no additional fields in the fit options structure.

If the Method field has the value SmoothingSpline, the SmoothingParam field is available to configure the smoothing parameter. Its value must be between 0 and 1. The default value depends on the data set.

If the Method field has the value LinearLeastSquares, the additional fields available in the fit options structure are listed in the table below.

If the Method field has the value NonlinearLeastSquares, the additional fields available in the fit options structure are listed in the table below.

Note   For the fields Upper, Lower, and StartPoint, the order of the entries in the vector value is the order of the coefficients returned by the coeffnames method. For example, if f = fittype('b*x^2+c*x+a'); coeffnames(f) ans = 'a' 'b' 'c' then setting options.StartPoint = [1 3 5]; assigns initial values to the coefficients as follows: a = 1, b = 3, c = 5. Note that this is not the order of the coefficients in the expression used to create f with fittype.

Example

Create the default fit options structure and set the option to center and scale the data before fitting:

options = fitoptions;
options.Normal = 'on';
options
options =
    Normalize: 'on'
    Exclude: [1x0 double]
    Weights: [1x0 double]
    Method: 'None'

Modifying the default fit options structure is useful when you want to set the Normalize, Exclude, or Weights fields, and then fit your data using the same options with different fitting methods. For example:

load census
f1 = fit(cdate,pop,'poly3',options);
f2 = fit(cdate,pop,'exp1',options);
f3 = fit(cdate,pop,'cubicsp',options);

Data-dependent fit options are returned in the third output argument of the fit function. For example:

[f,gof,out] = fit(cdate,pop,'smooth');
smoothparam = out.p
smoothparam =
    0.0089

The default smoothing parameter can be modified for a new fit:

options = fitoptions('Method','Smooth',...
                     'SmoothingParam',0.0098);
[f,gof,out] = fit(cdate,pop,'smooth',options);

Example

Create a noisy sum of two Gaussian peaks—one with a small width, and one with a large width:

a1 = 1; b1 = -1; c1 = 0.05;
a2 = 1; b2 = 1; c2 = 50;
x = (-10:0.02:10)';
gdata = a1*exp(-((x-b1)/c1).^2) + ...
        a2*exp(-((x-b2)/c2).^2) + ...
        0.1*(rand(size(x))-.5);
plot(x,gdata)

Fit the data using the two-term Gaussian library model:

f = fittype('gauss2');
gfit = fit(x,gdata,f)
gfit =
     General model Gauss2:
       gfit(x) =  a1*exp(-((x-b1)/c1)^2) + 
                  a2*exp(-((x-b2)/c2)^2)
     Coefficients (with 95% confidence bounds):
       a1 =    -0.05388  (-0.136, 0.02826)
       b1 =      -2.651  (-2.718, -2.584)
       c1 =     0.05373  (-0.04106, 0.1485)
       a2 =       1.012  (1.006, 1.018)
       b2 =      0.6703  (0.06681, 1.274)
       c2 =        41.2  (36.54, 45.85)

The algorithm is having difficulty, as indicated by the wide confidence intervals for some of the coefficients. To help the algorithm, we could specify lower bounds for the nonnegative amplitudes a1, a2 and widths c1, c2:

options = fitoptions('gauss2');
options.Lower = [0 -Inf 0 0 -Inf 0]; 

Recompute the fit with the bound constraints on the coefficients:

gfit = fit(x,gdata,ftype,options)
gfit =
     General model Gauss2:
       gfit(x) =  a1*exp(-((x-b1)/c1)^2) + 
                  a2*exp(-((x-b2)/c2)^2)
     Coefficients (with 95% confidence bounds):
       a1 =       1.003  (0.9641, 1.042)
       b1 =          -1  (-1.002, -0.9987)
       c1 =     0.04972  (0.04748, 0.05197)
       a2 =       1.002  (0.999, 1.004)
       b2 =       1.136  (0.725, 1.547)
       c2 =       48.89  (45.32, 52.47)

This is a much better fit. The fit can be further improved by assigning reasonable values to other fields in the fit options structure.

See Also

cflibhelp, fit, get, set

fit

fittype