Collection of vast amounts of information became feasible for many
businesses, organisations and institutions due to the availability of
inexpensive storage, communication and computer systems. In practice,
the collected information could be several hundreds of terabytes.
However, collection of such a large volume of information is worthless
for these organisations unless the data can be used for inferring some
useful knowledge that can be employed for improving the efficiency of
their systems and/or processes. In my talk I introduce the notion of
Interesting Patterns from data sets, which can be used for understanding
existing processes, or building efficient classifiers and intelligent
decision systems. Discovering these patterns is, however,
computationally a very hard problem in general. I will present efficient
ways of discovering some of these patterns, so that they can be used for
building intelligent decision systems. |