摘要
首先用中心化法对光谱数据进行预处理。建立极限学习机模型,其输入层神经元数为1 557,输出层神经元数为1,隐含层神经元数为10。模型得到的训练集均方根误差为0.103 4,预测集均方根误差为0.115 4。结果表明,极限学习机算法能够较准确地检测普洱茶中的游离氨基酸总量。
The spectrum data are pre-dealt with the mean centered method. The Extreme Learning Machine (EI,M) is established with 1 557 input neuron layers, 1 output neuron layer and 10 hidden neurons layers. The root mean square error of calibration set from ELM is 0. 103 4 where the root mean square error of prediction set (RMSEP) is 0. 115 4. The results show that the algorithm can precisely detect the free amino acid content in Purer tea.
出处
《长春工业大学学报》
CAS
2012年第3期269-273,共5页
Journal of Changchun University of Technology
基金
国家自然科学基金资助项目(30760103)
关键词
普洱茶
游离氨基酸总量
近红外光谱
极限学习机
Pu' er tea
free amino acid content
near infrared spectroscopy
Extreme LearningMachine (ELM).