Can I extract the fully-constructed model from a cfit object, not just the separate pieces of the formula?

2 次查看(过去 30 天)
I have a cfit object that I would like to get the full equation from, with parameter values plugged in. All I have been able to find in my searching so far is "coeffvalues" to get an array of parameter values and "formula" to get the general model form. For example, the cfit object is:
General model Exp1:
val(x) = a*exp(b*x)
Coefficients (with 95% confidence bounds):
a = 3695 (3307, 4083)
b = 9.448e-05 (8.636e-05, 0.0001026)
What I am looking to extract is:
val(x) = 3695*exp(0.00009448*x)
Is this possible?
  2 个评论
Steven Lord
Steven Lord 2020-9-2
Note that what you're looking to extract is unlikely to give the same results as evaluating the fit object itself. The values of the coefficients that are displayed may not be exactly the same as the values of the coefficients that are stored and used for computations.
Is a exactly 3695 or is it 3695.0000000001?

请先登录,再进行评论。

采纳的回答

Ayush Gupta
Ayush Gupta 2020-9-2
There is no direct way to obtain the equation with the substituted parameter values from the 'cfit' object in Curve Fitting Toolbox. However, there is a workaround, a string representation can be generated of the curve from a cfit object, ‘cfit_object’:
equation = formula(cfit_object); %Formula of fitted equation
parameters = coeffnames(cfit_object); %All the parameter names
values = coeffvalues(cfit_object); %All the parameter values
for i = 1:numel(parameters)
param = parameters{i};
l = length(param);
location = regexp(equation, param); %Location of the parameter within the string
while ~isempty(location)
%Substitute parameter value
equation = [equation(1:location-1) num2str(values(idx)) equation(location+l:end)];
location = regexp(equation, param);
end
end

更多回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Linear and Nonlinear Regression 的更多信息

产品


版本

R2020a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by