Abstrakt

Behavioural Model of Adult Obesity by Childhood Predictors using Crowd Sourcing

Mohd Zareef Affani, Uma N. Dulhare

Childhood obesity in one of the most serious public health challenges of 21st century. Child development has different stages, so it isn’t always easy to know when a child is obese or overweight. Child development refers to the biological, psychological and behavioural changes that occur between birth and end of adolescence. Effective tools are required to determine the behaviours earlier in life and find its influence on weight gain later in life. Crowdsourcing can be used as a tool to assess childhood predictors of adult overweight or adult obesity. In our model, it describe an approach to machine science by allowing the non domain experts to collectively calculate the known and unknown predictors and provide responses to those predictors, such that they are predictive of some behavioural outcome of interest. This was done by building a Web platform and allowing the user to respond to questions which will help to predict a behavioural outcome and it also allow the user to pose new questions. These results in a dynamically building up online survey, but the result of this cooperative behaviour leads to models that can predict the user’s outcomes based on their responses to the user-generated survey questions. In our approach we develop a site that will lead to models that can predict user’s body mass index. In our approach, it also covers several areas which are identified by earlier research, such as parenting styles, dieting and healthy lifestyle. The results indicate that Crowdsourcing can reproduce already existing hypotheses and also generate new ideas. Users were able to determine the predictors for higher BMI, such as low physical activity in their lifestyle.