Association rules in data with various time periods

Abstract

In this paper an Association Rules data mining technique is adopted to explore the co-movement between sector indices listed on the Warsaw Stock Exchange. The indices are related to the various sectors of the economy. Because of the different time ranges the various indices are traded, the special approach has been used. That allowed us to analyze data in a wide range of time. The results were compared to those obtained using the tradi-tional approach. We observed higher values of measures and smaller errors for a majority of rules.