取代芳烃类化合物是重要的化工产品,应用广泛,但大多数具有毒性。本研究根据定量结构-活性相关性(QSAR)原理,研究了38种取代芳烃对发光菌的急性毒性(EC_(50))与其分子结构之间的关系,还应用遗传算法筛选出5个与EC_(50)最为相关的描述符,并应用多元线性回归方法和支持向量机方法建立QSAR模型。结果表明,两种方法所见模型的复相关系数、留一法交互验证系数分别为0.988、0.979和0.991、0.981,对外部预测样本的复相关系数和外部测试集交互验证系数分别为0.913、0.904和0.924、0.906,与相关文献比较,所建QSAR模型均具有更好的预测能力和稳健性。 Substituted aromatic compounds are important chemical products, widely used, but most are toxic. In this study, based on the quantitative structure-activity relationship (QSAR) principle, the relationship between the acute toxicity (EC 50) of 38 substituted aromatic compounds and their molecular structures was studied. (50) the most relevant descriptors, and apply multiple linear regression and support vector machine method to establish QSAR model. The results show that the complex correlation coefficient and the one-way mutual verification coefficient of the two methods are 0.988, 0.979 and 0.991, 0.981 respectively. The correlation coefficients between the external prediction samples and the external test sets are 0.913, 0.904 And 0.924,0.906 respectively. Compared with the related literatures, the constructed QSAR models have better predictive ability and robustness.