分析化学
分析化學
분석화학
CHINESE JOURNAL OF ANALYTICAL CHEMISTRY
2009年
12期
1754-1758
,共5页
戴益民%李浔%梁波%杨道武%曹忠%黄可龙
戴益民%李潯%樑波%楊道武%曹忠%黃可龍
대익민%리심%량파%양도무%조충%황가룡
醇%碳-13核磁共振%化学位移%分子结构参数%定量结构-波谱关系
醇%碳-13覈磁共振%化學位移%分子結構參數%定量結構-波譜關繫
순%탄-13핵자공진%화학위이%분자결구삼수%정량결구-파보관계
Alcohol%~(13)C nuclear magnetic resonance%chemical shift%molecular structural parameter%quantitative structure-spectrum relationship
用新颖的原子拓扑矢量Y_C、原子平衡电负性q_e、结构信息参数[N_H~i(i=α,β)]和γ校正参数对63个无环饱和脂肪醇的局部化学微环境进行了结构表征,并对化合物~(13)C NMR化学位移进行了QSSR研究.采用偏最小二乘回归得到模型的复相关系数R和标准偏差S分别为0.9915和2.4827;对353个碳原子~(13)C NMR化学位移的实验值与计算值的平均绝对误差仅为2.01×10~(-6).同时,采用留分法(Leave-molecule-out)和外检验方法测试模型的内部稳定性和外部预测能力.与文献结果比较,本研究所用参数少,且计算简便.
用新穎的原子拓撲矢量Y_C、原子平衡電負性q_e、結構信息參數[N_H~i(i=α,β)]和γ校正參數對63箇無環飽和脂肪醇的跼部化學微環境進行瞭結構錶徵,併對化閤物~(13)C NMR化學位移進行瞭QSSR研究.採用偏最小二乘迴歸得到模型的複相關繫數R和標準偏差S分彆為0.9915和2.4827;對353箇碳原子~(13)C NMR化學位移的實驗值與計算值的平均絕對誤差僅為2.01×10~(-6).同時,採用留分法(Leave-molecule-out)和外檢驗方法測試模型的內部穩定性和外部預測能力.與文獻結果比較,本研究所用參數少,且計算簡便.
용신영적원자탁복시량Y_C、원자평형전부성q_e、결구신식삼수[N_H~i(i=α,β)]화γ교정삼수대63개무배포화지방순적국부화학미배경진행료결구표정,병대화합물~(13)C NMR화학위이진행료QSSR연구.채용편최소이승회귀득도모형적복상관계수R화표준편차S분별위0.9915화2.4827;대353개탄원자~(13)C NMR화학위이적실험치여계산치적평균절대오차부위2.01×10~(-6).동시,채용류분법(Leave-molecule-out)화외검험방법측시모형적내부은정성화외부예측능력.여문헌결과비교,본연구소용삼수소,차계산간편.
A newly developed topological vector of atom Y_C, equilibrium electro-negativity of atom q_e, molecular structural information parameter[N_H~i(i=α, β)] and γ calibration parameter were used to describe the local chemical microenvironment of 63 acyclic alcoholic compounds. Quantitative structural spectrum relationships (QSSR) was systematically made on relationship between ~(13)C NMR chemical shifts of 353 carbon atoms and their molecular structure descriptors. By partial least square regression(PLS), the statistical results indicated that the model correlation coefficient and standard error were 0.9915 and 2.4827, respectively. And the average absolute error was only 2.01×10~(-6) between the calculated and experimental chemical shifts for 353 carbon atoms. To validate the estimation stability for internal samples and the predictive capability for external samples of resulting models, leave-molecule-out(LMO) cross validation(CV) and external validation were performed. Compared with the reported result, not only the number of descriptors employed in this study was much fewer, but also the calculation was much easier.