中药寒热预测系统的建立

Construction of A Prediction System for Chinese Materia Medica Hot/Cold Properties

中药寒热性质是中药药性的重要特点之一,作者在本文着重介绍了构建的中药寒热性质预测系统,并对中药寒热药性的化学基础进行了探讨。作者建立的寒热预测系统对中药寒热药性的预测准确率,训练集达到83.3%,测试集达到81.0%。在此预测模型的基础上,提出了原子数、氢键受体数、电负性表面面积等5个化学描述符,用于解释中药寒热药性的化学基础。

Chinese Materia Medica（CMM） properties are characterized as the hot/cold properties, of which we discussed and analyzed chemical fundamentals and our prediction system for CMM was presented particularly. The predictions accuracy of this system are 83. 3% and 81% in training and test set respectively. Besides, five descriptors derived from the system such as atomic numbers, the number of H-bond aeeeptors and negative charged surface area, had been put forward to give help on interpretation the relationship between hot/cold properties of CMM and their chemical ingredients.