机构地区: 深圳大学化学与化工学院
出 处: 《火炸药学报》 2007年第3期9-15,共7页
摘 要: 运用神经网络模型,采用误差反向传播算法,对一系列芳香族多硝基化合物的密度进行了预测。结果表明,芳香族多硝基化合物的密度与其分子结构存在良好的相关性,选用分子结构描述码作为输入特征参数能取得较高的预估精度,预测结果的相对误差一般在±10%以内。 The densities of a series of aromatic polynitro compounds are predicted via an artificial neural network (ANN) based on erroneous reversed dissemination method. The results show a better correlation between the densities and molecular structures of aromatic polynitro compounds. Selecting molecular structure describers (MSD) as input characteristics parameters, a better predicted accuracy is obtained. The relative error between the predicted values and literature ones of the densities of aromatic polynitro compounds is within ±10%.
关 键 词: 结构化学 人工神经网络 密度预估 芳香族多硝基化合物
领 域: [化学工程]