论文部分内容阅读
针对常用的质量建模方法精度不高且难以给出预测区间,提出了基于小波相关向量机的产品质量模型.应用仿真数据和带钢热镀锌锌层质量的实际生产数据分别建立了小波相关向量机模型.结果表明,小波相关向量机方法与支持向量机及传统的相关向量机相比,具有更好的预测精度,而且给出了预测区间.多组带钢热镀锌锌层质量实际数据的相对预测误差的平均值为4·52%,为保证产品质量提供必要的决策支持和分析手段.
Aiming at the low precision of the commonly used methods of quality modeling and the difficulty in giving the prediction interval, a product quality model based on wavelet correlation vector machine is proposed. The wavelet correlation is established by using the simulation data and the actual production data of galvanized steel strip quality. Vector machine model.The results show that the wavelet correlation vector machine method has better prediction accuracy than the support vector machine and the traditional correlation vector machine and the prediction interval is given.The quality of the hot- The average relative prediction error of data is 4.52%, which provides the necessary decision support and analysis tools to ensure product quality.