目的 研究豨莶草炮制前后滋味的变化，建立生、制品豨莶草滋味判别模型。方法 通过电子舌测量12批生品豨莶草与12批炮制品豨莶草的酸、苦、涩、咸、甜味值，利用配对t检验、主成分分析（PCA）和线性判别因子分析（LDA）等方法研究豨莶草炮制前后各味道的变化情况。结果 通过配对t检验发现豨莶草炮制后，苦、咸味下降，酸、涩、甜没有明显变化；通过PCA可以区分生品与制品豨莶草；通过LDA可以建立生品、制品味道判别模型，并进行交叉验证，正确率达到100%；通过雷达图建立了豨莶草生品与制品的滋味指纹图谱。结论 电子舌能够对中药豨莶草的酸、苦、涩、咸、甜味值进行准确的辨识，结合统计方法，可以从数值上体现豨莶草炮制前后滋味的改变，并以此推测豨莶草炮制前后滋味的改变与功效的改变之间的关系。另外，建立的生、制品判别函数模型及雷达图指纹图谱可以用于生、制品豨莶草的判别。

Objective To study the taste of unprocessed and processed Siegesbeckiae puescens (SP) and establish a taste discrimination model. Methods The sour, bitter, astringent, salty, and sweet values of 12 batches of unprocessed and processed SP were measured by electronic tongue using a paired t-test, principal component analysis (PCA), and the linear discriminant factor analysis (LDA) to study the changes of the taste of unprocessed and processed SP. Results The paired t-test showed that the bitterness and saltiness of processed SP decreased; Astringency, sweetness, and sourness did not significantly change. The PCA can distinguish unprocessed and processed SP. The model of the taste of unprocessed and processed SP was established through the LDA, and conduct cross validation; Its correct rate reached 100%. The fingerprint of taste of unprocessed and processed SP was established through radar-graph. Conclusion Electronic tongue can identify sour, bitter, astringent, salty, and sweet values of traditional Chinese medicine accurately. It can reflect the numerical change of taste of unprocessed and processed SP combining with statistical method. Based on it, it can be inferred that the relationship between the change of taste and clinical efficacy before and after processing of SP. Besides, the discriminant model can be used to distinguish unprocessed and processed SP.

国家自然科学基金项目资助（81173550）