论文部分内容阅读
将自适应映射 (SOM)用于多环芳烃致癌性的分级。采用的输入参数为分子比表面积、代谢活性区及亲电活性区的中心碳原子离域能、分子中脱毒区总数。优化的网络参数包括网格数及网格形状、学习次数和学习率、邻居半径等。在最佳网络参数下 ,多环芳烃致癌性分类准确度大于 97%。
Adaptive Mapping (SOM) is used for the classification of carcinogenic PAHs. The input parameters used are the molecular specific surface area, the central carbon atom delocalization energy of the metabolically active region and the electrophilic active region, and the total number of drug-free regions in the molecule. Optimized network parameters include the number of grids and grid shape, the number of learning and learning rate, the radius of neighbors and so on. Under the optimal network parameters, the classification accuracy of PAHs is greater than 97%.