Weka: Way Cool

Written for STAT 503 after watching some videos about Weka.

These videos were a very nice introduction to the capabilities of Weka. In the first video, we learned how to classify using logistic regression, decision trees, neural networks and support vector machines using Weka. The Weka interface reminds me a lot of JMP, with all of the nested options and the pointing and clicking. This makes Weka very simple to use and to explore. I think that using Weka to explore classification problems is probably a very good way to start learning about classification, statistical learning, and data mining. Being able to interact with the data and the classifiers so easily and quickly is really beneficial to the learner.

In the videos from the Weka MOOC, we learned that a weka is a type of bird, which I thought was cute. I didn’t understand why they named it after a bird, though. But that’s not the most interesting thing from the MOOC videos at all. The “user classifier” option in Weka is very cool. I think that’s probably the most brilliant classification tool I’ve seen so far in any course. While it’s not the greatest classifier in the world, with an error rate in the example of 21% from a very experienced user, it’s an excellent tool for learning about classification. It is especially excellent for visualization of the process of classification. This is great for all those visual learners out there, including myself. Hopefully, I can get Weka running on my own computer and play around with it very soon!