基于SVM算法的中药材产地鉴别模型研究  

Study on Identification Model of Chinese Medicinal Herbs Based on Support Vector Machine Algorithm

作　　者：杨环瑜 陈莹 丘彤彤 张子荣 罗伟军 

机构地区：[1]湛江幼儿师范专科学校,广东 湛江 [2]湛江机电学校,广东 湛江

出　　处：《理论数学》2023年第2期364-374,共11页Pure Mathematics

摘　　要：中药材在大辞典中已记载有12,000多种,而不同类中药材又分布在众多的产地中,因此鉴别中药材产地的任务极为艰巨。本文以2021年“高教社杯”全国大学生数学建模竞赛E题“附件3”所提供的近红外光谱数据与中红外光谱数据为研究样本,先对所提供数据进行预处理、提取特征向量、降维处理,然后通过支持向量机(SVM)算法进行求解并运用评价指标在训练集上达到了83.8%的正确率,在测试集上达到98.7%的正确率。There are more than 12,000 kinds of traditional Chinese medicine recorded in the dictionary, and different kinds of traditional Chinese medicine are distributed in many places of origin, so the task of identifying the place of origin of traditional Chinese medicine is extremely difficult. In this paper, the near-infrared spectral data and mid infrared spectral data provided by the 2021 “Higher Education Society Cup” National Undergraduate Mathematical Modeling Contest E question “Ap-pendix 3” are taken as research samples. First, the data provided are preprocessed, feature vectors are extracted, and dimensions are reduced. Then, support vector machine (SVM) algorithm is used to solve the problem and evaluation indicators are used to achieve 83.8% accuracy in training sets and 98.7% accuracy in test sets.

关 键 词：中药材产地 支持向量机 红外光谱 机器学习 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]