论文部分内容阅读
目的利用近红外光谱技术建立蝙蝠拟青霉菌丝体中腺苷和多糖含量的定量分析模型。方法采用常规方法对蝙蝠蛾拟青霉菌丝体中腺苷和多糖含量进行测定,利用近红外光谱技术建立测定菌丝体中腺苷和多糖含量的相关模型,并通过蒙特卡罗偏最小二乘法(Monte Carlo Partial Least Square,MCPLS)和可移动窗口偏最小二乘法(Moving Window Partial Least Square,MWPLS)对模型进行优化。结果该模型校正集预测值和真实值间的相关系数(Rc)分别为0.9400和0.8781,预测均方根误差(Root Mean Square Error of Prediction Set,RMSEP)分别为0.5949和1.6617,校正均方根误差(Root Mean Square Error of Calibration Set,RMSEC)分别为0.5844和1.5572。结论该模型的稳健性、拟合度和预测能力均能达到令人满意的程度,该方法可以推广应用到其他发酵产品的检测。
Objective To establish a quantitative model of adenosine and polysaccharide content in Paecilomyces bats by near-infrared spectroscopy. Methods The contents of adenosine and polysaccharides in the mycelium of Paecilomyces militaris were determined by the conventional method. The correlation between adenosine and polysaccharides in mycelia was established by near infrared spectroscopy. The Monte Carlo Partial Least Square (MCPLS) and the Moving Window Partial Least Square (MWPLS) are used to optimize the model. Results The correlation coefficient (Rc) between predicted and true values of the model calibration set was 0.9400 and 0.8781 respectively. The Root Mean Square Error of Prediction Set (RMSEP) was 0.5949 and 1.6617 respectively. The root mean square error of correction (Root Mean Square Error of Calibration Set, RMSEC) were 0.5844 and 1.5572 respectively. Conclusion The robustness, fitness and predictive ability of this model can reach satisfactory levels. This method can be applied to the detection of other fermented products.