I'm working at classifying documents according to their content.
First I built a decision tree model that gives 90% of goods results.
Then I tried a TFIDF/SVC approach which also gives 90% of good results.
So now i'd like to combine both. My first thought was to add the prediction of TFIDF/SVC as a feature of the decision tree.
I saw this post about bagging/stacking/boosting. For me, adding the feature to the decision tree is equivalent to stacking. Is that correct ?