Yang Bao, Zhao Mouming, Jiang Yueming
South China Botanical Garden, Chinese Academy of Sciences, Guangzhou 510650, PR China.
College of Light Industry and Food Sciences, South China University of Technology, Guangzhou 510640, PR China.
Food Chem. 2008 Sep 15;110(2):294-300. doi: 10.1016/j.foodchem.2008.01.067. Epub 2008 Feb 8.
Various ultrasonic conditions were employed to prepare polysaccharides from longan fruit pericarp (PLFP) and the Lineweaver-Burk equation was then used to determine the effect of PLFP on inhibition of tyrosinase activity. This result showed that PLFP acted as a non-competitive inhibitor of tyrosinase. The highest slope was observed for ultrasonic extraction, followed by the hot-water extraction, suggesting that the ultrasonic treatment of PLFP increased the inhibition of tyrosinase activity. Furthermore, a multilayer feed-forward neural network trained with an error back-propagation algorithm was used to evaluate the effects of ultrasonic power, time and temperature on the slope value. The trained network gave a regression coefficient (R(2)) of 0.98 and a mean squared error (MSE) of 0.58, implying a good agreement between the predicted value and the actual value of the slope, and confirmed a good generalization of the network. Based on the artificial neural network-genetic algorithm, the optimal ultrasonic extraction conditions to obtain the highest slope value (154.1) were determined to be 120W, 12min and 57°C. Application of response surface plots showed the slope value as a function of every two factors under various ultrasonic extraction conditions, which can be observed directly. Therefore, the artificial neural network provided a model with high performance and indicated the non-linear nature of the relation between ultrasonic conditions and slope value.
采用不同的超声条件从龙眼果皮中制备多糖（PLFP），然后用Lineweaver-Burk方程确定PLFP对酪氨酸酶活性抑制的影响。结果表明，PLFP作为酪氨酸酶的非竞争性抑制剂。超声提取的斜率最高，其次是热水提取，这表明PLFP的超声处理增强了对酪氨酸酶活性的抑制作用。此外，使用误差反向传播算法训练的多层前馈神经网络来评估超声功率、时间和温度对斜率值的影响。训练后的网络给出的回归系数（R(2)）为0.98，均方误差（MSE）为0.58，这意味着预测值与斜率的实际值之间具有良好的一致性，并证实了网络具有良好的泛化能力。基于人工神经网络-遗传算法，确定获得最高斜率值（154.1）的最佳超声提取条件为120W、12分钟和57°C。响应面图的应用显示了在各种超声提取条件下斜率值作为任意两个因素的函数，这可以直接观察到。因此，人工神经网络提供了一个高性能的模型，并表明了超声条件与斜率值之间关系的非线性性质。
