论文部分内容阅读
为提高金属轧制预测模型中各参数数据的有效率,保证预测值的准确性,提出一种基于改进模糊C均值聚类算法的数据去噪方法。通过密度划分方法初始化聚类中心,设计了采用欧氏距离与夹角余弦结合而成的相似度指标的FCMD聚类算法,实现对参数预测所需数据的精确聚类;进而设计了一种双尺度噪声判别方法进行噪声去除。将该方法应用于实际板形预测模型进行实验,结果表明,基于FCMD聚类的去噪算法能有效提升样本数据的信噪比,降低均方根误差,提升轧制参数预测的准确度。
In order to improve the efficiency of each parameter data in metal rolling prediction model and ensure the accuracy of prediction, a data denoising method based on improved fuzzy C-means clustering algorithm is proposed. By using density partitioning method to initialize the clustering center, an FCMD clustering algorithm based on the similarity index formed by the Euclidean distance and the included cosine is designed to achieve the accurate clustering of the data needed for the parameter prediction. Then a double Scale noise discrimination method for noise removal. The method is applied to the actual flat prediction model. The experimental results show that the denoising algorithm based on FCMD clustering can effectively improve the signal-noise ratio of the sample data, reduce the root-mean-square error and improve the accuracy of rolling parameter prediction.