Bonding effect prediction of dry-cleaned composite fabric by principal component analysis and neural network

Eight principal components were obtained from the related parameters of the fabric and adhesive lining through principal component analysis,which were used as new variables.The BP neural network technology was adopted to construct a three-layer neural network model for prediction of the bonding effect of dry-cleaned composite fabric,and the model was trained using the vector and the algorithm for learning to adapt to new situations.The comparison of the predicted values and the experiment test values indicated that the prediction of bonding effect of dry-cleaned composite fabric by neural network is rather accurate and this testified in a certain extent that this method is practical.