Abstract:
Influence of items on some other items might not be the same as the association between these sets of items. Many tasks of data analysis are based on expressing influence of items on other items. In this paper, we introduce the notion of an overall influence of a set of items on another set of items. We also propose an extension to the notion of overall association between two items in a database. Using the notion of overall influence, we have designed two algorithms for influence analysis involving specific items in a database. As the number of databases increases on a yearly basis, we have adopted incremental approach in these algorithms. Experimental results are reported for both synthetic and real-world databases.
