Authors:
Subhagata Chattopadhyay
;
Pradeep Ray
and
Lesley Land
Affiliation:
APuHC, SISTM, Australian School of Business, University of New South Wales, Australia
Keyword(s):
Health data, Data mining, ANOVA, Regressions, QAR.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Knowledge Acquisition
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Symbolic Systems
Abstract:
Demographic health indicators such as crude birth rate, crude death rate, maternal mortality rate, infant mortality rate (IMR), Adult literacy rate and many others are usually considered measures of a country’s health status. These health indicators are often seen in an isolated manner rather than as a group of associated events. Conventional statistical techniques often fail to mine inter-relations among these indicators. This paper focuses on mining association-correlations among various demographic health indicators under child immunization program, skilled obstetric practice, and IMR using both statistical and Quantitative Association Rule (QAR) mining techniques. Relevant archived data from 10 countries located in the Asia-Pacific region are used for this study. Finally the paper concludes that association mining with QAR is more informative than that of statistical techniques. The reason may lie in its capability to generate the association rules using a 2-D grid-based flexible
approach. Finally it is concluded that such an approach could be pioneering for engineering the hidden knowledge among various other health indicators.
(More)