摘要
针对鹅膏真菌分类的问题,区别于传统人工分类方法操作比较复杂,分类速度慢的问题。从鹅膏属真菌特征形态数据入手,提出基于机器学习算法的一种分类模型。模型前期采用最值归一化的方法对数据进行预处理,其次以支持向量机作为分类器,选择线性函数作为核函数,同时使用交叉验证算法来寻找最优的惩罚系数C,通过混淆矩阵来分析数据分类处理中产生的误差,最后对试验中两种最优核函数的改进混合,使得模型的精准率在一定程度上达到最优。在分类模型中,将15个鹅膏属真菌的形态特征作为分类依据,并且给予相关属种的分类指标。在将数据进行相关要求划分后,将数据分为训练集和测试集,用训练集数据进行分类模型的建立,测试集数据验证其分类精准度。通过试验得出最后测试模型的精准度达到91.43%,训练模型的测试得分达到97.59%,在进行核函数的混合改进后测试模型的精准度以及训练模型的精准度都达到100%。
The problem of classification of Amanita Fungi is different from the traditional manual classification method, which is more complicated in operation and slow in classification. In this study, a classification model based on machine learning algorithm was proposed based on the characteristic morphology data of Amanita Fungi. In the early stage of the model, the data is preprocessed by the most value normalization method. Secondly, the support vector machine is used as the classifier, the linear function is selected as the kernel function, and the cross-validation algorithm is used to find the optimal penalty coefficient C. Analyze the errors generated in the data classification process, and finally improve the mixture of the two optimal kernel functions in the experiment, so that the accuracy of the model is optimal to some extent. In the classification model, the morphological characteristics of 15 Amanita Fungi were used as classification basis, and the classification indicators of related species were given. After dividing the data into relevant requirements, the data is divided into training set and test set, and the training set data is used to establish the classification model, and the test set data is used to verify the classification accuracy. Through experiments, the accuracy of the final test model reached 91.43%, and the test score of the training model reached 97.59%. After the hybrid improvement of the kernel function, the accuracy of the test model and the accuracy of the training model reached 100%.
作者
李卓识
陈晓旭
温长吉
王娜
林楠
田霞
Li Zhuoshi;Chen Xiaoxu;Wen Changji;Wang Na;Lin Nan;Tian Xia(College of Information Technology,Jilin Agricultural University,Changchun,130118,China;Key Laboratory of Engineering Bionics,Ministry of Education,Jilin University,Changchun,130025,China)
出处
《中国农机化学报》
北大核心
2020年第1期136-143,共8页
Journal of Chinese Agricultural Mechanization
基金
吉林省自然科学基金(20180101041JC)
吉林省教育厅“十三五”科学技术项目(JJKH20180659KJ、2016186).
关键词
鹅膏真菌
机器学习
支持向量机
混合核函数
分类模型
Amanita Fungi
machine learning
support vector machine
mixed kernel function
classification model
作者简介
第一作者:李卓识,男,1980年生,吉林长春人,博士研究生,副教授,研究方向为信息化农业、仿生科学与农业工程。E-mail:leezs643@sina.com;通讯作者:温长吉,男,1979年生,吉林吉林人,博士,副教授,研究方向为机器视觉与机器学习。E-mail:chenxiaoxu080628@163.com
