Hi professor,

You said the most complex part is the convert of the arrff file, and then we basically should do the same thing in easy.py.

But when we run the train method, we need the parameters g & c which looks like should be get from grid.py.

If the above is correct, then we should also do the things in grid.py OR find a good pair of g & c for car.data.arff and hard code them?

Hi Dr. sun,

1.  Can I assume the libsvm.jar is already referenced in the build path on your machine?

2.  Can the parameters change base on the dataset?  For example, if C-SVC is better for allelectronics.arff while nu-SVC is better for employees.arff, can I display different parameters for each dynamically? 

Thank you.