基于IGS聚类模型的中药材鉴别方法研究  

Research on Identification Method of Traditional Chinese Medicine Based on IGS Clustering Model

作　　者：窦婷 曹益兴 Dou Ting;Cao Yixing(Army Engineering University of PLA,Nanjing 210000)

机构地区：[1]中国人民解放军陆军工程大学,江苏南京210000

出　　处：《中阿科技论坛（中英文）》2024年第12期40-44,共5页China-Arab States Science and Technology Forum

摘　　要：中药材因其化学成分的差异性,具备独特的红外光谱特征,这些特征对于中药材品种的鉴定至关重要。在应用中红外光谱技术对中药材进行分类时,通常采用聚类算法与降维技术相结合的策略来处理光谱数据。但是,传统的聚类分析方法的结果容易受到聚类数目选择的影响,这限制了样本识别率的提升。针对这一挑战,文章采用了IGS聚类模型,有效地估计了425种中药材样本的最佳聚类数目,并获得了更为合理的聚类结果。这种方法不仅提高了中药材识别的准确性,也增强了识别过程的合理性。Chinese medicinal materials exhibit unique infrared spectral characteristics due to their chemical composition differences,which are crucial for the identification of Chinese medicinal material varieties.When applying infrared spectroscopy technology to classify traditional Chinese medicinal materials,a combination of clustering algorithms and dimensionality reduction techniques is usually used to process spectral data.However,the results of traditional clustering analysis methods are easily affected by the selection of the number of clusters,which limits the improvement of sample recognition rate.In response to this challenge,this article adopted the IGS clustering model to effectively estimate the optimal number of clusters for 425 samples of traditional Chinese medicine,and obtained more reasonable clustering results based on this.This method not only improves the accuracy of traditional Chinese medicine identification,but also enhances the rationality of the identification process.

关 键 词：中红外光谱技术 IGS模型 最佳聚类数 主成分分析 中药材鉴别 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]