文献类型：会议论文

英文题名：Identification of Chinese medicinal materials based on Support Vector Machine

作者：Hu, Zhengkun[1]; Liu, Xiaoyu[1]; Chen, Zhansheng[1]; Zhang, Puyuan[1]

第一作者：胡正坤

机构：[1] College of Applied Science and Technology, Beijing Union University, China

第一机构：北京联合大学应用科技学院

通讯机构：[1]College of Applied Science and Technology, Beijing Union University, China|[1141775]北京联合大学应用科技学院;[11417]北京联合大学;

会议论文集：Proceedings of 2024 International Conference on Cloud Computing and Big Data, ICCBD 2024

会议日期：July 26, 2024 - July 28, 2024

会议地点：Dali, China

语种：英文

外文关键词：Near infrared spectroscopy - Support vector machines

摘要：In the development and research process of traditional Chinese medicine, the identification and analysis of Chinese medicinal materials is an important link in quality control and inspection. The identification and quantitative analysis of effective ingredients for efficient identification and quality control is a difficult research challenge. The use of computer technology, statistical analysis, and mathematical modeling to extract features (peak position, peak intensity, and peak shape) of Chinese medicinal materials is a fast, accurate, and non-destructive identification technique for identifying the types and origins of Chinese medicinal materials. The SVM in Python was used for training, the mid-infrared spectral and near-infrared spectral data were selected as features, and the two kernel functions of Linear and ploy were used respectively, and the near-infrared spectral model was obtained through parameter tuning and mutual verification, with the accuracy of the training set being 99% and the accuracy of the test set being 98%. The accuracy of the model obtained from the mid-infrared spectrum was 100% in the training set and 91.8% in the test set. The data of all known categories of medicinal materials were selected for analysis, and the SVM was used for training to obtain a classification model of medicinal material categories, with an accuracy of 100% for the training set and 100% for the test set, and the classification model could be used to identify the medicinal materials in Annex 2 and obtain the corresponding categories. Secondly, according to the category of medicinal materials, SVM was used to train each type of medicinal materials, and the classification model of each medicinal material origin was obtained, and the accuracy of the training set was 93.8%, and the accuracy of the test set was 87.5%. The accuracy of the training set was 100%, and the accuracy of the test set was 44.4%, which was relatively low. C The accuracy of the training set is 100%, and the accuracy of the test set is 100%, and the identification results of the origin can be predicted according to the classification model of the three medicinal materials. ? 2024 ACM.