why SVM predicted rows more than the RF and Decision tree

Hi,

I run a prediction model and after dividing the data into 2 parts, training and testing, I run 3 different algorithms.

After running the SVM, RF and decision tree classification algorithm, when I look at the confusion matrix and the Predictor node of each algorithm, the number of tested rows is the same for RF and decision tree but more for SVM. What is the reason. I would appreciate if someone can explain.

Thank you.