From data mining to knowledge mining: Application to intelligent agents

The last decade, the computers world became a huge wave of data. Data mining tasks were invoked to tackle this problem in order to extract the interesting knowledge. The recent emergence of some data mining techniques provide also many interesting induction rules. So, it is judicious now to process these induction rules in order to extract some new strong patterns called meta-rules. This work explores this concept by proposing a new support for induction rules clustering and classification. The approach invokes k-means and k-nn algorithms to mine induction rules using new designed similarity measures and gravity center computation. The developed module have been implemented in the core of the cognitive agent, in order to speed up its reasoning. This new architecture called the Miner Intelligent Agent (MIA) is tested and evaluated on four public benchmarks that contain 25,000 rules, and finally it is compared to the classical one. As foreseeable, the MIA outperforms clearly the classical cognitive agent performances.