Incremental Dependency Aided Metapattern Generator

Abstract

Accumulation of data in electronic format is increasing at an exponential rate. Valuable transaction data will remain static and unexploited until analyzed to acquire knowledge. This work aims at enhancing a data mining technique called Metapattern generation. Domain knowledge is exploited in our approach and integrated with an existing Data Mining algorithm to generate interesting patterns that were not generated by the existing traditional techniques. Furthermore, new techniques are employed to enable the original algorithm to cope with incremental data.