摘要
旋光实验中,采用BP神经网络模型进行蔗糖浓度定标.在Matlab6.5环境下,借助NNTool完成训练、测试.对未知浓度蔗糖溶液进行预测,将预测结果与最小二乘法拟合值作比较,并讨论了隐含层神经元个数对BP神经网络预测精度的影响.
A BP neural-network model is adopted for calibration of sucrose concentrations in optical rotation experiment. In the Matlab 6.5 environment, neural-network is trained and tested by virtue of NNTool. Based on the model, unkown concentrations of sucrose solution are predicted. The results are compared with the values fitted by least-square method. And effects of hidden unit numbers on the prediction accuracy are discussed.
出处
《河南师范大学学报（自然科学版）》
CAS
CSCD
北大核心
2012年第6期56-58,共3页
Journal of Henan Normal University(Natural Science Edition)
基金
江西省自然科学基金项目(2010GQW0006)
赣南师范学院教学改革研究项目(JG-1035)
关键词
旋光实验
旋光度
神经网络
预测
optical rotation experiment
angle of optical rotation
neural-network
prediction
