机织物透气性预测的投影寻踪回归模型

Projection pursuit regression model for prediction of air permeability of woven fabrics

针对机织物透气性预测中存在非线性建模困难的问题,选择机织物总紧度、厚度、面密度及平均浮长等结构参数作为机织物透气性预测的影响因素,建立机织物透气性预测的投影寻踪回归模型。对模型训练样本的拟合值及检验样本的预测值以相对误差的均值及标准差为指标进行分析,并与BP神经网络及多元线性回归模型进行对比。结果表明,投影寻踪回归模型的拟合及预测精度均优于BP神经网络及多元线性回归模型,且在训练样本较少的情况下,投影寻踪回归模型仍有较高的预测精度和较强的泛化能力,可为机织物透气性预测提供一种新的方法。

In view of the difficulty in nonlinear modeling for prediction of woven fabric permeability,a projection pursuit regression（PPR） model for prediction of air permeability of woven fabrics was established using the structural parameters such as the total tightness,thickness,and weight per square meter and average float as factors affecting the prediction of woven fabric permeability.The fitted values of tested samples and the predicted values of trained samples were analyzed with the means and standard deviations of relative error as the indicators and were compared with the results of BP neural network and multiple linear regression model.The results showed that the PPR model fitting and prediction accuracy was better than those of BP neural network and multiple linear regression model.In the case of less trained samples,the PPR model still had relatively high prediction accuracy and good generalization ability,providing a novel approach to the prediction of woven fabric permeability.