ISSN : 2319-7323

INTERNATIONAL JOURNAL OF COMPUTER SCIENCE ENGINEERING

ABSTRACT

Title	:	An Effective Mining of Exception Class Association Rules from Medical Datasets
Authors	:	Abdualkareem Ali Al-Tapan, Basheer Mohamad Al-Maqaleh
Keywords	:	Association Rule Mining; Knowledge Discovery; Medical Data Mining; Class Association Rule; Exception Class Association Rule.
Issue Date	:	Aug 2017
Abstract	:	Data mining techniques can be applied on medical data to improve quality of decisions in medical field, and to enhance clinician performance. Class Association Rule (CAR) integrates the techniques of association rule mining and classification rules mining. CAR mining is able to find a lot of rules, some important rules are still not discovered. These rules are called Exception Class Association Rules (ECARs), which have high confidence and low support. The goal of ECAR is to discover rare and low support itemsets to generate useful rules from these itemsets and meet the class label (target attribute). Also, ECARs are considered very important by domain's expert in medical field, which lead to improve the performance of analysis and diagnosis. These type of rules cannot be identified easily using traditional algorithms as they focus in frequent itemsets discovery and don't depend on target attribute for diagnosis. In this paper, an effective algorithm is proposed to mine two types of CARs: Precise Class Association Rules (PCARs) and Approximate Class Association Rules (ACARs), and to generate ECAR for each ACAR from medical datasets. The experimental results shows the effectiveness of the proposed algorithm.
Page(s)	:	191-198
ISSN	:	2319-7323
Source	:	Vol. 6, No.8