基于生物电阻抗的人体内脏脂肪预测方法

Visceral fat estimation method based on bioelectrical impedance technology

为了得到比CT、MRI等传统内脏脂肪测量法成本更加低廉、安全性高、辐射更低的方法,进行基于生物电阻抗技术的内脏脂肪组织的预测模型的研究.提出一种新的内脏脂肪预测方法,通过先验知识复杂化特征属性,使用赤池信息量准则(AIC)简化模型得到基本特征属性,并在此基础上通过复杂和多样的信号分析和建模,运用模糊逻辑的思想,使用多线性回归模型取代了单一的线性模型,使得预测值精确性得以提高.同时对训练数据集进行详细的数据分析及验证,为今后如何进一步提高模型预测精准度做出了判断.实验结果表明,该方法的预测结果好于单一线性回归模型.

Bioelectrical impedance based visceral fat estimation methods were researched,in order to get a less expensive,lower radiative and more safe method comparing with traditional ones like X-ray CT and MRI.A new visceral fat estimation method was presented,which applied priori knowledge to carry out the complication,followed by the simplification based on Akaike information criteria（AIC）.With complicated and diversified signal analysis and modeling,fuzzy logic was used to create multi-linear regression models to achieve high accuracy.Furthermore,more detailed data analysis was done on the training data to judge how to improve the model in the future.Experimental results show that this method has high estimation accuracy and stability.