论文部分内容阅读
利用线性判别分析和概率神经网络,建立了预测中草药有效成分利尿性与其分子结构参数之间的 QSAR 模型．概率神经网络分类结果好,训练集、交互检验集和测试集的分类正确率均可达到100％．本文所用的概率神经网络结构简单、易于调试,研究工作进一步明确了分子利尿性与其结构参数之间的关系,有助于利尿药物的选择与合成．
Using linear discriminant analysis and probabilistic neural networks, a QSAR model was developed to predict the diureticity of its active ingredients and its molecular structure parameters. Probabilistic neural network classification results are good, and the classification accuracy of training set, cross-examination set and test set can reach 100%. The probabilistic neural network used in this paper has a simple structure and is easy to debug. The research work further clarifies the relationship between molecular diuresis and its structural parameters, which contributes to the selection and synthesis of diuretic drugs.