中文摘要：

采用主成分分析法（PCA）对样本数据集进行预处理，将得到的新样本数据集输入支持向量机（SVM），籍助均匀设计（UD），构建几丁质酶氨基酸组成和最适温度的数学模型．当径向基核函数的3个参数，惩罚系数C为10，ε为0．5，γ为5时，模型对温度拟合的平均绝对百分比误差为5．06％，预测的平均绝对误差为1．83℃，说明具有良好的预测效果且优于神经网络的预测结果．

英文摘要：

The principal component analysis was applied to the data processing in training sets, the new principal components were then used as input data of support vector machine modle. A prediction model for optimum temperature of chitinase was established based on uniform design. When the regularized constant C, ε and γ were 10, 0.5 and 5, respectively, the calculated temperature fitted the reported optimum temperatUre of chitinase very well and the mean absolute per cent error （MAPEs） was 5.06%. At the same time, the predicted temperature fitted the reported optimum temperature well and the mean absolute error （MAE） was 1.83 ℃. It was superior in fittings and predictions compared to the model based on back propagation neural network.