高盐稀态酱油二次沉淀预测模型的构建和验证

为构建高盐稀态酱油二次沉淀预测模型，该研究对比分析了与酱油二次沉淀形成相关的理化指标，包括pH值及铁离子、亚铁离子、多酚、多糖、氯化钠、乙醇、谷氨酸和温度等指标。利用SPSS软件分析了酱油二次沉淀生成量与各指标之间的相关性，并以此为基础建立与二次沉淀相关的多元线性回归预测模型。结果表明，高盐稀态酱油二次沉淀生成量（Y，g/L）与pH值（X1）极显著相关（p<0.01）、与铁离子（X2，mg/L）、多酚（X4，g/100 mL）和氯化钠含量（X6，g/100 mL）均显著相关（p<0.05），并成功构建了它们之间的多元线性回归方程。验证实验结果表明，高盐稀态酱油二次沉淀的多元线性回归方程预测值与存放3个月后的实测值具有良好的相关性（R2=0.8517）。该预测模型有助于提前发现会产生严重二次沉淀的酱油，避免其流入市场和给企业造成经济和声誉损失，具有重要的应用价值。

Establishment and Validation of Prediction Model for the Secondary Precipitate in High-salt Diluted-state Soy Sauce

In order to establish the prediction model of secondary precipitate of high-salt dilute-state soy sauce, the physicochemical indexes related to the formation of secondary precipitate of soy sauce were comparatively analyzed, including pH value, iron ion, ferrous ion, polyphenol, polysaccharide, sodium chloride, ethanol, glutamic acid and temperature. SPSS software was used to analyze the correlation between the amount of secondary precipitate and each index, a multiple linear regression prediction model based on the correlation analyses was established. The results showed that the amount of secondary precipitate (Y, g/L) was significantly correlated with pH value (X1) (p<0.01), iron ion (X2, mg/L), polyphenol (X4, g/100 mL) and sodium chloride contents (X6, g/100 mL) (p<0.05), and the multiple linear regression equation amongst them was successfully established. The results showed that there was a good correlation between the predicted value of multiple linear regression equation and the measured value after 3 months storage (R2=0.8517). The prediction model is helpful to find the soy sauce that will produce serious secondary precipitate in advance, and avoid its inflow into the market and cause economic and reputation losses to enterprises, thus it has important application value.