Weakly Supervised Learning Algorithm to Eliminate Irrelevant Association Rules in Large Knowledge Bases

Abstract

Large knowledge bases construction and population have being a very explored in the past few years. Many techniques were developed in order to accomplish this purpose. Association rule mining algorithms can also be used to help populating these knowledge bases. Nevertheless, analyzing the amount of association rules extracted can be a hard task, spending a lot of time. In this way, this article presents a weakly supervised learning association rule mining algorithm to eliminate irrelevant association rules extracted. The proposed method uses rules already discovered in past iterations and prunes off those with the same pattern. Experiments showed that the new technique can reduce and eliminate the amount of rules in about 60\%, decreasing the effort spent on evaluation them.