How to plot feature importance on x-axis and variable names on y-axis?

4 vues (au cours des 30 derniers jours)
Devendra
Devendra le 8 Août 2023
Commenté : Devendra le 9 Août 2023
I wish to plot variables names on y-axis and feature importance on x-axis in Random Forest Regression. Also the variables on y-axis should be in descending order according to their feature importance on x-axis. I request you to kindly suggest me how to do it. D S Chahar
  1 commentaire
Devendra
Devendra le 8 Août 2023
I want to generate the feature importance plot similar to attached image. I request you all to please suggest me how to create it in matlab. Many thanks in advance.
Devendra

Connectez-vous pour commenter.

Réponse acceptée

dpb
dpb le 8 Août 2023
Modifié(e) : dpb le 8 Août 2023
See barh and particularly <the example>. Use the second optional output from sort to get the order of the variables when you sort them to put in descending order to use with xticklabels to label the y-axis tick values. Or, you could convert to categorical variable type for the x-value (as ordinal to keep the sorted order) and will get the labels automagically.
nfacts=8;
features="f"+[0:nfacts-1].';
F=randi([20,125],size(features));
[F,ix]=sort(F);
hB=barh(F);
yticklabels(features(ix))
xlabel('F Score'), ylabel('Features')
hTxt=text(hB.YEndPoints,hB.XEndPoints," "+string(F),'horizontal','left','vertical','middle');
NOTA BENE: with the horiontal bar, the x,y postions for text() are y,x on the plot axes nomenclature; that's easy to confuse and the text doesn't show up or is somewhere bizarre.
  3 commentaires
dpb
dpb le 9 Août 2023
"...keep the values ... up to two decimal points only"
Just replace
" "+string(F)
generating the string to write with
compose(" %0.2f",F)
The olden way was with num2str
Devendra
Devendra le 9 Août 2023
Thanks a lot. It worked.
Devendra

Connectez-vous pour commenter.

Plus de réponses (0)

Catégories

En savoir plus sur MATLAB dans Help Center et File Exchange

Produits


Version

R2023a

Community Treasure Hunt

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

Start Hunting!

Translated by