Using Interval Association Rules to Identify Dubious Data Values

Abstract

A hard-to-catch erroneous data is one whose value looks perfectly legitimate. Yet, if we examine this value in conjunction with other attribute values, the value appear questionable. Detecting such dubious values is a major problem in data cleaning. This paper presents a framework to automatically detect dubious data values in the datasets. Data is first pre-processed by data smoothing and mapping. Next, interval association rules are generated which involved data partitioning via clustering before the rules are generated using an Apriori algorithm. Finally, these rules are used to identify data values that fall outside the expected intervals. Experiment results show that the proposed framework is able to accurately and efficiently dubious values in large datasets.