中文摘要：

以701种有机化合物的土壤吸附系数作为数据集,选取594种有机化合物作为训练集,剩余107种作为测试集.根据训练集化合物建立土壤吸附系数Koc与辛醇/水分配系数Kow的线性和非线性模型,应用平均残差（AE）,平均绝对残差（AAE）和均方根误差（RMSE）来检验模型的预测能力,模型具有良好的预测能力.同时,比较不同类型的化合物的实测值与预测值,发现部分同系物的预测值与实测值存在系统的偏差,这些偏差主要是由吸附机理,溶解度,水解作用,挥发作用,实验误差等原因造成,这些因素均会对土壤吸附系数的预测产生影响.

英文摘要：

The hydrophobic parameter represented by the octanol/water partition coefficient （Kow） is commonly used to predict the soil sorption coefficient （Koo）. In the present paper, soil sorption data for 701 compounds were analyzed. The results show that lgKoo is linearly related to lgKow for the compounds with lgKow in the range of 0. 5-7.5 and non-linearly related to lgKow for the compounds in a wide range of lgKow. A non-linear model was developed between lgKoc and lgKow for a wide range of compounds in the training set. The models were validated terms of average error （AE）, average absolute error （AAE） and root-mean squared error （RMSE） by using an external test set with 107 compounds. Systemic predictive deviations were been observed for some class-specific compounds. The reasons for systemic deviations may be attributed to the difference of sorption mechanism for hydrophilic compounds, low solubility for highly hydrophobic compounds, hydrolysis of esters in solution, volatilization for volatile compounds and highly experimental errors for compounds with extremely high or low sorption coefficient.