定量药物设计中病态矩阵理化指标剔除法

REJECTION OF PHYSICO-CHEMICAL INDICES FOR PATHOLOGICAL MATRIX IN A QUANTITATIVE DRUG DESIGN

在定量药物设计中,各理化指标X_i及生物活性Y之间的相关矩阵往往呈病态,求得的回归方程不稳定。本文绘出一种寻求稳定方程的方法:选择一个适当的临界值α(0<α<1),找出大于或等于临界值α的相关系我及所涉及的指标,剔除其中部分指标。将剩下的指标作逐步回归,从而得到了稳定的回归方程。

In a quantitative drug design, correlation coefficient matrix of physico-chemical indices X_i is often pathological when there indices are highy correlated. Theregression equation of biological activity Y about there indices X_i obtained in this case isnot stable. In this paper, a method is given to obtain a stable regression equation:giving acritical valne α and finding out the indices among which correlation coefficient is not lessthen α, The following are the rules to reject some of them. If X_i and X_j is highly correlated(|r_(ij)|≥α, r_(ij) is the correlation coefficient of X_i and X_j), and |r_(in)|>|r_(jn)|(r_(in) and r_(jn) are correlation coefficients of X_i and X_j about Y, respectively), then the index X_j is rejected, other-wise, X_i is rejected. Stable equation can be obtained by stepwise regression with the re-maining indices.