AI learning with decision trees
Horses and Riders
© Lead Image © Outdoorsman, fotolia.com
Decision trees help machines emulate human behavior. Thanks to artificial intelligence, programs can autonomously build decision trees by learning from existing real-life data. Can an AI system determine who was operating a particular vehicle by analyzing driving style?
In the 1950s, TV stations were broadcasting howlers that probably wouldn't interest anyone today. One was "What's My Line?" with presenter John Daly, in which a blindfolded team of four panelists had to guess the identity of a star guest using only yes or no questions.
A popular strategy was to pose general questions first ("Are you male?"), in order to ask more specific questions in the final round, until the panelists began to close in on the guest star and their identity was finally revealed.
In machine learning, programmers employ similar techniques to teach computers to imitate learned behaviors. As a somewhat contrived example, take the behavior of an AND gate (Table 1), which always shows the 0 value at output until a 1 is applied to both inputs. The gate is normally implemented by binary operators, but for the purpose of this investigation, I'll be using the decision tree in Figure 1.
[...]
Buy this article as PDF
(incl. VAT)
Buy Linux Magazine
US / Canada
UK / Australia