SEECS Colloquium

The extraction of logic from data, as part of data mining approaches, is becoming increasingly important. We describe a method that reliably and without tuning or parameter selection produces the desired logic functions plus probability distributions of their accuracy.

One option of the method allows enforcement of prior logic conditions for the learned formulas. Another option concerns the situation where data are obtainable at some cost. For that case, one can construct formulas that classify records at minimum cost.

We discuss some applications in intelligent systems such as medical diagnosis, natural language processing, and traffic control.

K. T. is Professor of Computer Science in the Erik Jonsson School of Engineering and Computer Science of the University of Texas at Dallas. Current research concerns computational logic including optimization, and intelligent computer systems. Ongoing projects cover integration of logic computation into programming languages, learning logic and data mining, natural language processing, learning algorithms, automated construction of expert systems, medical diagnosis, and traffic control.