IceCube: Efficient Targeted Mining in Data Cubes

Provided by: Indian Institute of Science
Topic: Data Management
Format: PDF
In this paper, the authors address the problem of Targeted Association Rule Mining (TARM) over complex multi-dimensional market-basket data cubes. Their goal here is to extend traditional association rule mining to capture targeted rules that apply to specific customer segments. Thus, the minimum support for rules in a particular segment is now a percentage of the number of transactions that map to that segment, rather than the total number of transactions over all the segments. To model this scenario, they consider multidimensional market-basket data in which each transaction has, in addition to the set of purchased items, ancillary dimension attributes associated with it.

Find By Topic