摘要
本文利用主成分提取-线性判别分析(PCA-LDA)模型对多环芳烃(Polycyclic Aromatic Hydrocarbons,PAHs)的致癌性进行分类,与致癌性有关的多环芳烃的表面积、代谢活性区域中心碳原子的离域能、亲电活性区域中心碳原子的离域能以及分子脱毒区的总数四个参数作为模型的输入,用已知致癌性的67个样本作为训练集建立PCA-LDA模型,对10个预测集样本的致癌性进行预测,结果表明:致癌性按高(h)、低(l)、非(n)分类时预测准确率达100%。
This paper adopts principal component analysis-linear discriminant analysis (PCA-LDA) to classify the carcinogenicity of polycyclic aromatic hydrocarbons. Four carcinogenic-related parameters such as the total surface area (TSA) of the molecule of polycyclic aromatic hydrocarbons, the delocalization energy of the center carbon atom in the metabolic active region (△E1), the delocalization energy of the center carbon atom in the electro-philic active region (△E2), and the number of de-toxic regions in the molecule (Nd) are used as inputs for the model. 67 carcinogenic samples are used as training set to build PCA-LDA model to predict the carcinogenicity of 10 prediction samples. The result shows that when the carcinogenicity is classified on high, low, and no level, the predictive accuracy reaches 100%.
出处
《分析科学学报》
CAS
CSCD
2007年第6期717-719,共3页
Journal of Analytical Science
基金
国家自然科学基金(No.60571055)
南通市科技项目基金(K2006007)
关键词
主成分分析
线性判别分析
多环芳烃
致癌性
Principal component analysis(PCA)
Linear discriminant analysis(LDA)
Polycyclic aromatic hydrocarbons(PAHs)
Carcinogenicity