Integrating User's Domain Knowledge with Association Rule Mining

TL;DR

This paper introduces a modified Apriori algorithm that integrates user domain knowledge to enhance the efficiency of association rule mining, reducing computational time and space by focusing on user-specified attributes.

Contribution

It proposes a novel variation of Apriori that incorporates user preferences, enabling more targeted and faster association rule discovery.

Findings

Incorporating user preferences improves search efficiency.

The modified algorithms outperform standard Apriori in speed and space.

User-guided attribute selection reduces computational resources.

Abstract

This paper presents a variation of Apriori algorithm that includes the role of domain expert to guide and speed up the overall knowledge discovery task. Usually, the user is interested in finding relationships between certain attributes instead of the whole dataset. Moreover, he can help the mining algorithm to select the target database which in turn takes less time to find the desired association rules. Variants of the standard Apriori and Interactive Apriori algorithms have been run on artificial datasets. The results show that incorporating user's preference in selection of target attribute helps to search the association rules efficiently both in terms of space and time.