目的研究多元线性回归法、逐步回归法和标准偏回归系数法3种关键工艺参数辨识方法的特点。方法以黄芩水提液酸沉碱溶工艺关键参数辨识为例比较3种方法。结果 3种方法所得结果基本一致,即关键参数为药液质量分数、酸沉pH值、碱溶pH值和静置温度。结论 3种方法的阈值选择都有主观性;多元线性回归法思路最简单,但辨识关键工艺参数不如逐步回归法灵敏;标准偏回归系数法的优势在于能综合参数对多个指标的影响,并且易于加权计算。 Objective To study the characteristics of three kinds of key process parameters identification methods: multiple linear regression method, stepwise regression method and standard partial regression regression coefficient method. Methods Scutellaria water acid soda alkali solution process of identification of key parameters as an example to compare three kinds of methods. Results The results obtained by the three methods were basically the same, that is, the key parameters were liquid mass fraction, acid precipitation pH value, alkali solution pH value and standing temperature. Conclusion The threshold selection of the three methods is subjective. The method of multiple linear regression is the simplest, but the identification of key process parameters is less sensitive than that of stepwise regression. The advantage of standard partial regression coefficient method is that it can comprehensively influence the influence of multiple parameters, Easy to weight calculation.