| 机器学习在近红外光谱法判别鲍鱼品种研究中的应用 |
| |
| 引用本文： | 高婧娴,黄扬明,雷春丽,闫红,闵顺耕,熊艳梅,闵志勇.机器学习在近红外光谱法判别鲍鱼品种研究中的应用[J].中国农业大学学报,2018,23(9):166-170. |
| |
| 作者姓名： | 高婧娴 黄扬明 雷春丽 闫红 闵顺耕 熊艳梅 闵志勇 |
| |
| 作者单位： | 中国农业大学理学院;莆田学院环境与生物工程学院 |
| |
| 基金项目： | 国家自然科学基金项目（31301685）；莆田市科技计划区域重点项目（2015N1002）；福建省科技计划区域重大项目（2009N3002） |
| |
| 摘 要： | 为解决市场上鲍鱼产品缺乏科学分类方法的问题,利用近红外光谱分析技术结合机器学习方法对鲍鱼快速分类进行研究,使用MicroNIRTM1700便携式近红外光谱仪采集3种鲍鱼,即绿盘鲍(25只)、红壳鲍(31只)、皱纹盘鲍(35只)的光谱数据,采用CART算法建立鲍鱼分类决策树模型,以模型对测试集样本的预测准确率衡量决策树模型优劣,分裂策略为在每个节点处选择Gini不纯度最大的方式进行分裂,通过交叉验证控制决策树深度。结果表明,对训练集180条光谱建立模型,采用5折交叉验证,模型准确率为90.00%,对测试集93条光谱的预测准确率为90.32%。本研究方法可以很好地区分绿盘鲍、红壳鲍和皱纹盘鲍,满足鲍鱼现场快速分类的需求。
|
| 关 键 词： | 近红外光谱 鲍鱼 水产品 机器学习 分类回归树 |
| 收稿时间： | 2017/10/30 0:00:00 |
Discrimination of abalone (sub) species basing on near-infrared spectroscopy and machine learning |
| |
| GAO Jingxian,HUANG Yangming,LEI Chunli,YAN Hong,MIN Shungeng,XIONG Yanmei and MIN Zhiyong.Discrimination of abalone (sub) species basing on near-infrared spectroscopy and machine learning[J].Journal of China Agricultural University,2018,23(9):166-170. |
| |
| Authors: | GAO Jingxian HUANG Yangming LEI Chunli YAN Hong MIN Shungeng XIONG Yanmei and MIN Zhiyong |
| |
| Institution: | College of Science, China Agricultural University, Beijing 100193, China,College of Science, China Agricultural University, Beijing 100193, China,College of Science, China Agricultural University, Beijing 100193, China,College of Science, China Agricultural University, Beijing 100193, China,College of Science, China Agricultural University, Beijing 100193, China,College of Science, China Agricultural University, Beijing 100193, China and College of Environmental and Biological Engineering, Putian University, Putian 351100, China |
| |
| Abstract: | To study the rapid classification of abalone (sub) species, near-infrared spectroscopy combined with machine learning method was used. The spectra of three (sub) species abalone samples were obtained by portable near-infrared spectrometer MicroNIRTM1700. The spectra was divided into training set and testing set, which were 180 and 93 spectra, respectively. The CART method was applied to build a decision tree model and its criterion was Gini impurity. Cross validation was used in the model to control the depth of decision tree model. The accuracy rate of the training set was 90.00%. The final accuracy rate of the testing set reached 90.32%. The combination of NIRS and chemometric method was proposed in this study as a fast and new method for the classification of different abalone (sub) species. |
| |
| Keywords: | near infrared spectroscopy abalone aquatic product machine learning classification and regression tree |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《中国农业大学学报》浏览原始摘要信息 |
| 点击此处可从《中国农业大学学报》下载免费的PDF全文 |
|
