Implementation of background knowledge and properties induced by fuzzy confirmation measures in apriori algorithm

Abstract

This work contributes to so-called association analysis. Its goal is to search for dependencies (called associations) between attributes in large scale data sets. Recently the authors theoretically studied some properties of fuzzy confirmation measures and possible application of background (resp. expert) knowledge into associations mining process. In this work we implement our recent results into well-known Apriori algorithm. Despite of the fact that the presented algorithm allows us to mine linguistic associations, i.e., associations interpretable in natural language, basic ideas of this algorithm can be easily extended to less specific model of fuzzy sets. © 2013 Springer-Verlag Berlin Heidelberg.