Parameter Identification and Selection for Childhood Obesity Prediction Using Data Mining

Abstract

The accurate identification and selection of useful parameters for childhood obesity prediction are very important. This study aims to identify childhood obesity prediction parameters for children in Malaysia, and presents the methods used to identify and select the parameters from the children's attributes, lifestyle, family, and environment. The study comprises four stages: risk factor review, data collection, parameter identifications and selection, and evaluation. Base on the results, 19 parameters were identified. The accuracy of childhood obesity prediction using the proposed parameters was 21% greater compared to a set of parameters used in a previous study.