Calculating the % of records correctly identified by Random Forests in R




I am currently trying to implement ensemble methods for one of the Kaggle competitions(Digit Recognizer).
I plan to use Random Forest,Neural Net and SVM and them combine them.
What I now want is to get a prediction % of the data,I mean how many records have been correctly classified.
So if RandomForest gives - 70% , Neurals- 80 % and SVM about 60% I will combine the results.
How do I go about doing this in R??