作者: Milan Vrbaški , Rade Doroslovački , Aleksandar Kupusinac , Edita Stokić , Dragan Ivetić
DOI: 10.1007/S12652-019-01374-3
关键词:
摘要: Lipid profile usually includes levels of total cholesterol (TCH), low density lipoprotein (LDL), high (HDL) and triglycerides (TG), all which require a blood test. Using advances in machine learning relationship between lipid obesity, model that predicts without using any laboratory results can be developed used clinical diagnosis. The causal obesity is well known—TCH, LDL TG show an increase, while HDL decreased obese persons. In this paper we are artificial neural networks (ANN) to estimate the values non-lab electronic health record data some measures obesity. ANN inputs gender, age, systolic diastolic pressures, single or combination multiple parameters, include body mass index, saggital abdominal diameter height ratio, waist ratio fat percentage. Study shows presented solution suitable for prediction TCH (with accuracy 81.89%), 79.29%) 81.23%), not 44.48%).