Analysing Sri Lankan Lifestyles with Data Mining: Two Case Studies of Education and Health

Abstract

There are no adequate researches in lifestyle data analysis of Sri Lanka. Existing works are not capable of handling big data systematically and, not efficient in disclosing the latent factors in lifestyles. This research has used predictive and descriptive mining techniques to analyse HIES dataset related to two cases in education and health using WEKA and SPSS. The design consists of a classification and a clustering task. Classification finds the factors and their relationships associated with no-schooling and dropouts. Clustering explores the relationship between chronic diseases and family income. Our analysis reveals the significance of child labor and religion among the influences such as age, district and parental education related to school dropout and no-schooling. Other case study discloses that citizens who have got low income comparatively suffer from many chronic diseases. The important patterns recognized through the research can be used by the government policy makers to setup policies.