论文部分内容阅读
改进了BP神经网络的学习速率算法 ,并运用改进后的BP网络 ,通过分子结构参数 ,同时对烷基吡啶在非极性固定液SE 30和极性固定液PEG 1 50 0柱上的保留指数进行预测 ,取得了较好的结果。
The algorithm of learning rate of BP neural network is improved. By using the improved BP network, the retention index of alkylpyridine on non-polar fixative SE 30 and polar immobilization PEG 1 50 0 is improved by molecular structure parameters Forecasting, and achieved good results.