智能网络系统低匹配度数据深度挖掘算法研究

Research on deep mining algorithm for low matching data in intelligent network system

常规的挖掘算法在处理智能网络系统中的低匹配度数据时,因迭代次数过多造成运算节点上的数据挖掘完成度较低.为此,提出智能网络系统低匹配度数据深度挖掘算法研究.经过数据变换和数据规约处理数据,将数据转换成更适合深度挖掘的格式,最大限度地精简数据量,利用神经网络模型更新挖掘规则,同时将数据结构中冗余连接与节点去掉,采用I/O输入输出方式循环计算完成数据深度挖掘.测试结果表明:相同的测试条件下,与常规的挖掘算法相比,提出的低匹配度数据深度挖掘算法在不同的运算节点上的挖掘完成度更高,始终保持在85%~99%之间,适合应用在智能网络系统低匹配度数据深度挖掘中.

Conventional mining algorithms deal with low-matching data in intelligent network systems,because the number of iterations leads to a low degree of completion of data mining on computing nodes.Therefore,a deep mining algorithm for low matching data in intelligent network system is proposed.Transform the data into a format more suitable for deep mining to process the data through data transformation and data reduction.This method can minimize the amount of data.Mining rules are updated using neural network modules,and redundant connections and nodes are removed from the data structure.Data deep mining is completed using I/O input and output method cyclic calculation method.Experimental results show that the experimental results show that compared with the conventional mining algorithms in the same test conditions,the low matching degree data deep mining algorithm proposed in this paper has higher mining completion on different computing nodes,which is always between 85%and 99%.It is suitable for the deep mining of low-matching data in intelligent network system.