文献类型：期刊文献

中文题名：基于回归算法与决策树思想的秦艽产地划分

英文题名：Distribution of Gentiana Macrophylla Based on Regression Algorithm and Decision Tree

作者：杨友[1];包彩银[1];丁跃武[1];李四海[1]

第一作者：杨友

机构：[1]甘肃中医药大学信息工程学院,甘肃兰州730000

第一机构：甘肃中医药大学信息工程学院（教育技术中心）

年份：2023

卷号：19

期号：2

起止页码：41

中文期刊名：亚太传统医药

外文期刊名：Asia-Pacific Traditional Medicine

基金：甘肃省科技计划(21JR1RA272)。

语种：中文

中文关键词：秦艽;产地;多元线性回归;偏最小二乘回归;支持向量机回归;岭回归;决策树近红外光谱

外文关键词：Gentiana Macrophylla;Origin;Multiple Linear Regression;Partial Least Squares Regression;Support Vector Regression;Ridge Regression;Decision Tree;Near Infrared Spectroscopy

摘要：目的:药材产地对药材品质优劣影响很大,不同产地秦艽品质差异较大,使用秦艽样本近红外光谱数据构建模型,实现秦艽样本产地划分。方法:将多元线性回归算法(MLR)、偏最小二乘回归算法(PLS)、支持向量回归算法(SVR)和岭回归算法(RR)与决策树思想相结合对样本进行分类。结果:秦艽数据集包含207个样本,其中107个样本来自甘肃玛曲县,100个秦艽样本来自甘肃临洮县,模型的误判率在0.3%左右。结论:将回归算法与决策树思想相结合的方法可以提高模型的性能,使得算法既拥有回归算法强大的拟合能力,又拥有决策树的分类能力。在秦艽样本数据上得到了很好的分类效果。
Objective:The origin of medicinal materials has a great influence on the quality of medicinal materials.The quality of Gentiana macrophylla in different origins varies greatly.The near-infrared spectral data of Qingang samples is used to build a model to achieve the division of Qingang samples.Methods:Samples were classified by combining Multiple Linear Regression(MLR),Partial Least Squares(PLS),Support Vector Regression(SVR) and Ridge Regression(RR) with decision tree ideas.Results:The Gentiana macrophylla dataset contained 207 samples, of which 107 samples were from Maqu County, Gansu Province, and 100 Samples from Lintao County, Gansu Province.The false positive rate of the model was about 0.3%.Conclusion:The classification algorithm that combines regression algorithm with decision tree idea can accurately classify the data.The algorithm has both the strong fitting ability of the regression algorithm and the classification ability of the decision tree.In the Gentiana macrophylla sample data, it has obtained a good classification effect.