Predicting Obesity from Grocery Data: A Conceptual Process Framework

Abstract

In upholding the Islamic way of life, effort to seek for moderation can be in the form of obesity prevention. Obesity is becoming the future burden of nations and actions have been taken to curb the problem of obesity. Most nations predict obesity based on the national past trend using data from population-based health surveys which are costly. Alternative method now points to data analytics which have created a productivity growth wave in business organization through its predictive and prescriptive capabilities. Exploratory works on the use of data analytics have extended into areas such as healthcare in the effort to improve the quality of life. To offer alternative methods of obesity prediction, we present our early work on applying data analytics in predicting obesity based on dietary pattern derived from the retailing transactional data of grocery shopping instead of using health-related data collected from periodic population–based survey. We described the framework development and later proposed a process framework for predicting obesity from grocery data. The first part of the framework describes the process of deriving dietary intake patterns from grocery data and the second phase describes the obesity prediction process that utilizes data mining tools. This paper also reviews the data mining techniques that had been applied in obesity prediction in previous researches.