Using WEKA with separate training and test data sets and employing cross validation

I have confused myself regards training and testing a classifier using WEKA.

I have a dataset that consists of approximately 120 samples (timestamped signal data features). These samples are divided into 6 repetitions of a human hand movement (i.e. the movement is repeated 6 times). The paper I am reading states the following: Reps 1, 3, 4 and 6 are used for training while Reps 2 and 5 are used for testing.

Therefore, I have put the training data in one file and the testing data in another file. I also have to use cross validation on this data, which I know WEKA provides.

Now, to run the classifier with this data I thought the procedure would be as follows: 1. Select the Preprocess tab. 2. Open the file with the training data reps in 3. Select the Classify tab and choose a classifier 4. Select the "Cross-validation" option and run the classifier on this data 5. Then use the "Supplied test set" option to select the file containing the testing data reps 6. Run the classifier again

But I suspect this is incorrect.

Can anyone please explain how I go about creating the model using the training data with cross validation and then using the test data on it?