Turning Information Into Knowledge



The new technology of "machine learning" will soon help the agricultural sector make better use of its information, thanks to a team of researchers at Waikato University and the support of the Foundation for Research, Science and Technology.

Machine learning enables a computer program to automatically analyse a large body of data and decide which information is most relevant. This crystallised information can then be used to help people make decisions faster and more accurately.

"One of the central problems of the information age is dealing with the enormous explosion in the amount of raw information that is available," said Professor Ian Witten, director of the project. "Machine learning has the potential to sift through this mass of information and convert it into knowledge that people can use." Professor Witten went on to explain that machine learning has so far been used only on small problems under well controlled conditions. "Our goal is to bring the technology out of the laboratory and provide solutions that can make a difference to people."

To this end, the research team is incorporating a number of machine learning techniques into a system called WEKA, for "Waikato Environment for Knowledge Analysis," an experimental facility for machine learning that is drawing international attention. With WEKA, a specialist in a particular field can use machine learning to derive useful knowledge from databases that are far too large to be analysed by hand. It has already been used to investigate what information dairy farmers use in deciding which cows to keep in their herds. Other problems to be tackled include analysing milk production and identifying the factors that govern the cyclical patterns of possum population.

As part of their work, the Waikato researchers are formulating new ways to evaluate machine learning techniques, and they are developing new machine learning methods of their own.

"The knowledge locked up in databases is a major economic force," said Professor Witten. "If it can be mined, refined, and recorded, it has the potential to contribute to the economy in many different ways. This is a direct way of increasing the country's assets."

Contact: Professor Ian H. Witten, Department of Computer Science, Waikato University

Phone: (07) 856-2889
Fax: (07) 838-4155

ihw@cs.waikato.ac.nz



If you have any comments about these pages then please contact: mlwebmaster@cs.waikato.ac.nz