Yarn Unevenness Prediction using Generalized Regression Neural Network

<p>This study aimed to propose a method predict yarn unevenness grounded on the generalized regression neural network and traditional model further improve prediction accuracy. The was constructed. Under this model, three-layer network, four-layer five-layer were designed. Finally, Python used for training simulation. parameters three models data made consistent ensure comparability of results. results showed that using average relative error cut down 0.87% compared with network. Compared performance not much different, but running speed increased by 46.05%. reduced 0.57%, mean square 0.98%, he root 4.76%, 74.70%.</p> <p> </p>