广州化工
廣州化工
엄주화공
GUANGZHOU CHEMICAL INDUSTRY AND TECHNOLOGY
2015年
16期
6-7,10
,共3页
曹红翠%孙海霞%保英莲
曹紅翠%孫海霞%保英蓮
조홍취%손해하%보영련
多溴联苯醚( PBDEs )%支持向量机%QSPR
多溴聯苯醚( PBDEs )%支持嚮量機%QSPR
다추련분미( PBDEs )%지지향량궤%QSPR
polybrominated diphenyl ethers ( PBDEs )%support vector regression%quantities structure property relationship ( QSPR)
应用支持向量机方法研究了19种PBDEs类化合物的P L 和分子结构的定量关系，结果发现，采用支持向量机方法可以实现使用较少样本数据建模，并达到较好预测结果的目的。支持向量机方法的预测结果远优于线性溶解能关系法的预测结果。用支持向量机回归方法对PL 进行了预测，准确度非常高，所得模型的平均相对误差为0.12262。
應用支持嚮量機方法研究瞭19種PBDEs類化閤物的P L 和分子結構的定量關繫，結果髮現，採用支持嚮量機方法可以實現使用較少樣本數據建模，併達到較好預測結果的目的。支持嚮量機方法的預測結果遠優于線性溶解能關繫法的預測結果。用支持嚮量機迴歸方法對PL 進行瞭預測，準確度非常高，所得模型的平均相對誤差為0.12262。
응용지지향량궤방법연구료19충PBDEs류화합물적P L 화분자결구적정량관계，결과발현，채용지지향량궤방법가이실현사용교소양본수거건모，병체도교호예측결과적목적。지지향량궤방법적예측결과원우우선성용해능관계법적예측결과。용지지향량궤회귀방법대PL 진행료예측，준학도비상고，소득모형적평균상대오차위0.12262。
Support vector machine was used to study PL and molecular structure relationship of 20 polybrominated diphenyl ethers.The results showed that fewer samples were used to set up a model and we could get a good out put moreover, SVMs'result was better than linear salvation energy relationships'.The study also showed that estimated PL of PBDEs by support vector regression method had a very good high accuracy , the average relative error of models were 0.12262 , respectively.
