Interestingness measure and multi support value for quantitative association rules

Data Mining is a field that has been used extensively in recent years to obtain meaningful and useful information from data accumulated in digital environments. Association Rule is one of the Data Mining techniques that are used to reveal relationships between data. The association rule is first discussed in 1993 by Agrawal et al. The support and confidence values used in association rules are the two most important parameters of this method. However, determining the value of support in quantitative data sets can be problematic. Failure to select the minimum support value causes some interesting and valuable rules in the dataset to either not be produced or to generate a large number of unnecessary rules that are not valuable. In this thesis, a new approach is proposed which uses multiple support values to overcome the problem of selecting support value on quantitative data sets. This proposed approach has been applied on two different sample data sets and the results obtained are explained in detail.