More Data Mining with Weka

11 April 2014

The University of Waikato is now offering an advanced MOOC on data mining.

More Data Mining with Weka will commence in late April and run for five weeks. It follows on from the entry-level MOOC, Data Mining with Weka, which the university is currently running for a second time.

Course convenor Professor Ian Witten says this advanced MOOC will provide a deeper account of data mining tools and techniques.

“Again the emphasis is on principles and practical data mining using Weka. Students will work with multimillion-instance datasets, classify text, experiment with clustering, association rules, neural networks, and much more,” he says. Enrolments are already open, and Professor Witten says that students should have completed the first MOOC, or have equivalent knowledge of the subject before signing up.

Open source software

The open source software Weka was developed at Waikato University and is widely used around the world — it’s downloaded more than once a minute, 24/7.  

“I have no idea who downloads it but I do know that anyone working with data needs to know about data mining,” says Professor Witten. “People need to understand the principles, but they don’t need the ‘dirty details’ and I think that’s what makes Weka so appealing.”

The first time the introductory MOOC was offered, 6500 students registered and 1050 completed successfully “which I understand is pretty good for MOOCs, which often have only a 5% completion rate,” Professor Witten says.

Great participation

There are 3500 students participating in the current offering. “We look at statistics, the principles of algorithms, different representations for data and its evaluation – it’s all the right stuff on how to apply data mining to answer a specific question. But you do need to know what question you’re trying to answer.”

The new MOOC contains 30 lessons (six a week), each comprising a short YouTube video and practical activities with Weka to reinforce learning. The aim is the same: to equip students to use advanced techniques of data mining on their own datasets – and understand what it is that they are doing. Designed for more advanced students, the practical activities are more extensive and challenging than before. “This is definitely learning by doing, not learning by listening,” says Professor Witten.

The video component of More Data Mining with Weka, like the introductory course, will also be offered on Youku, the Chinese version of YouTube, captioned in Chinese. China is one of the largest countries in the world for Weka downloads.

Latest stories