最佳投影识别法用于1-(1H-1,2,4-三唑-1-基)-2-(2,4-二氟苯基)-3-取代-2-丙醇及其衍生物抗真菌活性的分子筛选

Optimal Map Recognition Method Applied to Molecular Sieving of Antifungal Activities of 1- (1H-1, 2, 4-Triazole-1-Y1 )-2- (2, 4-Difluorophenyl)-3- Substituted-2-Propanols and Their Reduced Products

提出了一种新的模式识别方法──最佳投影识别法，该方法可自动从若干隐含的模式识别投影图中筛选出最佳的一个，并建立在原始特征空间中优类样本分布范围的边界方程．将该方法应用于 １－（１Ｈ－１，２唑－１－基）－２－（２，４－二氟苯基）－３－取代－２－丙及其衍生物的抗真菌活性的分子筛选，得到了建模所用的最佳模式识别投影图，建立了在由五个结构参数构筑的原始特征空间中该类化合物高抗菌活性样本分布范围的判据，并用以该类化合物高抗真菌活性分子的筛选。

A new pattern recognition method called optimal map recognition (OMR) method is proposed to select the optimal projection map among various pattern recognition maps. The inequalities describing the boundary conditions of optimal samples in original space can be established by OMR method. The OMR method has been successfully applied to computer - aided molecular sieving of antifungal activities of 1 -(1H - 1, 2, 4 - triazole - 1 - y1 )- 2 -(2, 4 - difluorophenyl)- 3 substituted - 2 - propanols and their reduced products. The new compounds with probably higher anti - fungicidal activities are determined by the inequalities describing the boundary conditions of samples with high antifungal activities in original space.