I run SMO and Naive Bayes for the same data set under Weka Experiment Environment. For SMO, I have 116.547 seconds for the train set and 19.865 seconds for the test set. For Naive Bayes, I have 80.665 seconds for the train set and 699.594 seconds for the test set. I wonder if these results are significant since Naive Bayes is known as a fast classifier.
Could anyone explain me these results?
How many instances do you have? Nowadays computers are pretty fast. Most probably, You will see drastic differences in speed in large datasets.
According to comment, you have 4k instance. so it is not so much. I used weka on KDD99 dataset with 4.000.000 instances. Training times measured in hours.