a computer-aided diagnosis system for the prediction of overweighs using life style factors, socio-economic status and family history of disorders in children

abstract background: obesity represents one of the most important nutritional problems worldwide.obesity in childhood can cause variety of health issues such as orthopedic, neurological, pulmonary and gastroenterological disorders in the future, although no side effects were reported from malignant obesity during childhood. in this paper, we presented a computer-aided diagnosis system to predict the obesity based on input features obtained from the life style and other factors of the subjects. methods: the total number of 9795 subjects (49.17% boy) aged 6 to 18 years taken from the caspian iv study participated in this study. the input parameters of the proposed system were taken from the dietary habit, physical activity, family history, social economic status, and other features. then, the obesity was predicted using the data mining and artificial intelligence techniques. feature selection (fs) methods were also used to improve the performance of the proposed system. the performance of the diagnosis system was assessed based on the hold-out validation framework . findings: the performance of the classifications method has been validated by hold-out cross validation. among the different classification techniques tested, svm with fs showed the best performance. the accuracy and precision of this method were 63.3% and 83.7%, respectively. some features such as age, physical activities, birth feeding and family history of diabetes mellitus detected as the most effectivefactors with obesity in both gender. conclusion: designing of an intelligentdiagnosis system with the input parameter such as life-style, socioeconomic status and genetic information can help predict obesity in children to modify their life-style to improve theirqualityof life in the future. a web-based version of this intelligent system can easily provide the obesity prediction facilities for the families at home. key words: classification; obesity; data mining; medical diagnosis system; artificial intelligence