目的 预测亚甲蓝及其衍生物的最大吸收峰，构建最大吸收峰与结构之间的构效关系（QSAR）模型。方法 利用收集到的18个亚甲蓝及其衍生物的结构和最大吸收峰数据，采用怀卡托智能分析环境（WEKA）的方法对数据进行归一化，建模和预测。结果 通过交叉验证和外部测试集对建立的拟合模型进行检测，得到相关系数R值分别为0.8和0.9。结论 通过WEKA的方法成功建立了亚甲蓝及其衍生物的结构和吸收峰的QSAR模型。该模型可为此类化合物开发为光敏剂提供依据，为某些有潜质的化合物进行结构改造做前期准备。

Objective To estimate the absorption peak of methylene blue and its derivatives and construct a quantitative structure-activity relationship (QSAR) model for the absorption peak and structure. Methods Structure and absorption peak data of methylene blue and 17 derivatives were collected. The Waikato Environment for Knowledge Analysis (WEKA) method was used to normalize the data and conduct the modeling and prediction. Results The constructed fitting model was tested via cross validation and an external test set. The values of derived correlation coefficient R of the model from two tests were 0.8 and 0.9. Conclusion A QSAR model for the structure and absorption peak of methylene blue and its derivatives has been successfully constructed via WEKA method. This model can provide evidence for developing these compounds into photosensitizer and transforming structures for some potential compounds.