Integration of Interactivity In Apriori Approach For Discovering Association Rules Hidden In the Target Dataset |

ABSTRACT:

Association rule mining provides valuable information interms of significant correlations between different attributes’ values thatmight not be evident at the first glance in large datasets. The experimentalpart of this work has demonstrated benefits of integration of interactivity inApriori approach for discovering association rules hidden in the targetdataset. The interactive algorithm for discovering association rules starts byasking user’s requirement with respect to attributes to be included in thesearch. Since the dataset has one class attribute that determines the patient class (LIVE or DIE),the clinicians are interested in finding rules that determine the value ofpatient class (LIVE or DIE). In addition to attribute specification, the user suppliesthe minimum support and confidence threshold, the twoparameters required by Apriori algorithm. In the experimental runs, minimum support and confidence threshold have been fixedat 15% and 80%, respectively