BioBodyComp: A Machine Learning Approach for Estimation of Percentage Body Fat

Bio Body composition (BBC) analysis describes and assesses the human body in various components such as total body water, lean muscle mass, fat mass, bone skeletal mass, and bone density. Excessive fat mass in the human body is associated with ill health and related to obesity, a phenotype of body composition. Obesity is a serious medical condition in which non-essential fat is accumulated along with a decrease in lean muscle mass. Body Mass Index (BMI), an equation, has been used for a very long time as a predictor of body fatness and obesity. As a predictor of obesity and based on height and weight, BMI is unable to explain and calculate the percentage body fat (PBF). BMI, which is based on only two anthropometric measurements, also misclassified obesity in many cases because it is not age and gender specific. Two people with the same height and weight i.e., BMI can have different PBF. An athlete who has more muscle mass than fatmass can also be misclassified as obese. BMI is a Rough Guide it cannot be used as an assessment tool for PBF. All these facts indicate that there is a need to develop a less complex technique to predict PBF and other body composition components. In this study, we have developed a normative and data-driven prediction model for structural body composition phenotype to predict PBF. The developed predictive model is based on less expensive and simple tomeasure anthropometric measurements such as age, gender, height, waist, hips, and weight.