小波神经网络测定复方苯甲酸酊中主要成分的含量  

Determination of compound benzoic acid tincture by wavelet neural network

作　　者：张成燕[1] 毛应明[1] 王学松[1] 张辉辉[1] 赵蕤[1] 

机构地区：[1]淮海工学院化学工程学院,江苏连云港222005

出　　处：《计算机与应用化学》2011年第10期1299-1301,共3页Computers and Applied Chemistry

基　　金：国家自然科学基金资助项目(20977040).

摘　　要：依据小波函数的非线性逼近能力和神经网络的自学习能力,用Morlet小波基函数代替传统的神经网络传递函数Sigmoid函数,克服BP神经网络易陷入局部极小等缺陷,建立了小波神经网络算法,用于紫外分光光度法不经分离同时测定复方苯甲酸酊中的苯甲酸和水杨酸。苯甲酸和水杨酸的平均加标回收率和RSD分别为101.1%,1.7%;102.1%,1.8%。结果表明,该方法不需要物理化学分离,分析速度快,精确度较高,适用于该制剂的含量测定。Based on the non-linear approximation ability of wavelet and the self-learning characteristic of neural network. With the Morlet wavelet function instead of the traditional transfer sigmoid function of neural network to overcome the BP neural network defects of be fall in local minimal, wavelet neural network model was establish to determine quickly benzoic acid and salicylic acid of Compound Benzoic Acid Tincture without pretreatment. For benzoic acid and salicylic acid, the average recovery and RSD was 101.1%, 1.7% and 102.1, 1.8%. The results showed that this method is more simple and it is not required to do physical or chemical separation, the speed of analysis is high and the accuracy is quite good.

关 键 词：苯甲酸 水杨酸 小波神经网络 复方苯甲酸酊 

分 类 号：TQ015.9[化学工程] TP391.9[自动化与计算机技术—计算机应用技术]