Combining two classifications for ensemble learning in R




I went through this post on ensemble learning in R and tried to implement something similar in R using

new_predict <- (knn_predict+randomForestPredict)/2 but it is giving an error saying:

I understand that this is because I cannot combine models predicting factors in this way.
How do I then achieve combining the results from KNN and Random Forests in R?
Can someone please help me with this!!


Alternatively, one way for ensembling in classification is to sum up the probability of predictions instead of summing up the classes which are factors. Hope this helps!


If you have multiple outputs already, generated by your KNN and RandomForest, you can average them with something very easy like which is the majority (for each individual / row) or something more advanced like weight each factor with the accuracy of each model.