LinearModel.fit results to array

7 vues (au cours des 30 derniers jours)
OoM le 26 Août 2013
Hi,
I would like to ask about results from LinearModel.fit function. I used the function and it ran smoothly. However, I could not extract the result properly.
I have a 100 set of X and Y. What I would like to have is a array showing coefficient and R-square of all. I basically try to do a loop and have LinearModel.fit function inside. My problem is I could not write out coefficient from mdl.Coefficients and r-square from mdl.Rsquared.Ordinary value to save into my array.
I crated array 100 x 1 to save R-square value from each set. To sum up, how can I save the values of results from each set to array?
Thank you.
0 commentairesAfficher -2 commentaires plus anciensMasquer -2 commentaires plus anciens

Connectez-vous pour commenter.

Réponses (3)

Leah le 26 Août 2013
I was able to do it, i made a dataset mid to hold everything
md = LinearModel.fit()
mid.Intercept=md.Coefficients.Estimate(1);
for c=2:length(md.CoefficientNames)
mid.(md.CoefficientNames{c})=md.Coefficients.Estimate(c);
end
2 commentairesAfficher AucuneMasquer Aucune
OoM le 26 Août 2013
Thank you very much.
OoM le 26 Août 2013
gave me
mid =
Rsq: 0.8690
It is in struct format. Any way to transform it to array
Rsq = [0.8690]

Connectez-vous pour commenter.

the cyclist le 26 Août 2013
mdl.Coefficients
is a variable of type dataset.
mdl.Coefficients.Estimate
will be an array of doubles, which I think is what you need.
2 commentairesAfficher AucuneMasquer Aucune
OoM le 26 Août 2013
I tried to call only mdl.Coefficients.Estimate. It does not work.
the cyclist le 27 Août 2013
Does calling this code give you a 5x1 double for mdl.Coefficients.Estimate?
X = ingredients; % predictor variables
y = heat; % response
mdl = LinearModel.fit(X,y)
mdl.Coefficients.Estimate

Connectez-vous pour commenter.

OoM le 26 Août 2013
I am able to get R-squared from
rsq = mdl.Rsquared.Ordinary
rsq =
0.8690
and only coefficient from
Coefficient = mdl.Coefficients.Estimate(:,1)
Coefficient =
1.0e+05 *
1.4231
-0.0012
0 commentairesAfficher -2 commentaires plus anciensMasquer -2 commentaires plus anciens

Connectez-vous pour commenter.

Catégories

En savoir plus sur Descriptive Statistics dans Help Center et File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by