Why don't they link to the paper? This article is really crappy and doesn't explain what they did at all. For example, why diddn't they mention probabilist networks, which dominates this field. And I doubt that 240 data points is enough to write a paper.