遗传算法在近红外无创伤人体血糖浓度测量基础研究中的应用  

Application of Genetic Algorithms in Fundamental Study of Non-invasive Measurement of Human Blood Glucose Concentration with Near Infrared Spectroscopy

作　　者：王宏[1] 李庆波[2] 刘则毅[1] 徐可欣[2] 

机构地区：[1]天津大学理学院数学系,天津300072 [2]天津大学理学院精密测试技术及仪器国家重点实验室,天津300072

出　　处：《分析化学》2002年第7期779-783,共5页Chinese Journal of Analytical Chemistry

基　　金：教育部科学技术研究重点项目

摘　　要：遗传算法 (GA)应用在偏最小二乘法 (PLS)校正模型的波长优化选择中具有显著的效果。将遗传算法作为模块循环运行 ,能更快达到最优解 ,有效提高测量精度 ,减少建模所用波长数。本文将该方法应用于无创伤人体血糖浓度光学检测的基础研究中 ,验证实验所用样品为 :①葡萄糖水溶液 ;②包含牛血红蛋白和白蛋白的葡萄糖水溶液 ;③人血中的血浆 (含葡萄糖 )。结果表明 :建模的波长个数可分别减少 88%、86 %、85 % ;预测标准偏差 (RMSEP)分别减少 5 6 %、6 4 %、6 3%。Genetic algorithms is an effective method in wavelength selection applied in building multivariate calibration model based on partial least squares regression. If genetic algorithm is run repeatedly as a block, the optimal solution is obtained faster, the numbers of wavelengths used to build calibration model is further reduced, the prediction precision is further improved. This method was applied to fundamental study of Non-Invasive of measurement of human blood glucose concentration with spectroscopy. The experiments tested in this method are (1) glucose in aqueous matrix, (2) glucose in aqueous matrix containing bovine serum albumin and heucoglobin and (3) human serum containing glucose. The result shows that the numbers of wavelengths for building the models can reduce by 88%, 86% and 85% respectively, while the root mean square error of prediction reduces by 56%, 64% and 63% respectively. It is instructive for the further study of the theory of non-Invasive measurement of human blood glucose with spectroscopy.

关 键 词：近红外光谱 无创伤检测 人体 血糖浓度 遗传算法 波长选择 偏最小二乘 

分 类 号：R311[医药卫生—基础医学]