基于混合累积模式匹配的云数据特征分区融合算法

Cloud Data Characteristics Partition Fusion Algorithm Based on Mixed Accumulation Pattern Matching

在无线传感器网络云计算环境下,传统的数据融合方法,对混合累积特征未进行滤波后置检测,无法实现模式匹配,数据融合误差控制不可控制。提出混合累积模式匹配的云数据特征分区融合算法,根据两云间云滴的取小取大后的比值描述云间相似程度,得到云数据信息熵融合特征的功率谱幅度,采用簇内数据相异粒度寻优法得到云数据的熵融合特征提取最优化的约束条件,对特征进行混合累积模式匹配,对云数据的分区特征混合累积模式匹配滤波后置处理,控制云数据分区特征融合误差,实现算法改进。仿真结果表明,该算法提高特征空间增益,其精度高、实现简单等优良特性,性能优于传统算法。

In the wireless sensor network (WSN) in cloud computing environment, the traditional method of data fusion filtering characteristics of mixed accumulation is not to the rear of detection, unable to realize pattern matching, data fusion control error control. Proposed hybrid cloud data accumulation pattern matching feature partition fusion algorithm, according to two clouds of cloud droplets take little take the ratio of the big describe the similarity degree between cloud, cloud data are features of power spectrum amplitude information entropy fusion, cluster data in different particle size optimization method is used to get the cloud data fusion feature extraction of entropy optimization constraints, the characteristics of hybrid cumulative pattern matching, and characteristics of cloud data partition mixed rear accumulation patterns matched filtering processing, error control cloud data partitioning feature fusion, implementation algorithm is improved. The simulation results show that the algorithm improve the characteristics of spatial gain, its fine properties such as high precision, simple implementation, performance is superior to the traditional algorithm.