论文部分内容阅读
数据融合算法能够实现对海量数据的整合和特征提取,以便形成更为清晰、可靠的数据,满足不同用户需求,但传统基于BP神经网络的数据融合算存在局部最优及泛化能力差的问题,本文引入了一种无监督学习技术自动编码器,并将其与分簇协议相结合衍生出了新型数据融合算法SAEMAD,最终经过实验对比,在同等条件下,该算法较BPNDA算法具有更好的数据特征提取优势。
Data fusion algorithm can realize the integration of large amounts of data and feature extraction in order to form more clear and reliable data to meet different user needs, but the traditional data fusion based on BP neural network has the problem of local optimization and poor generalization ability , This paper introduces an unsupervised learning technology automatic encoder and combines it with the clustering protocol to derive a new data fusion algorithm SAEMAD. Finally, through experimental comparison, under the same conditions, the algorithm has better performance than the BPNDA algorithm Data feature extraction advantages.