多数据链网络终端信息有效存储方法仿真

Simulation of Effective Storage Method For Terminal Information of Multiple Data Link Networks

针对当前网络终端信息存储的空间占用严重和耗时长的问题,提出基于关联维聚类的多数据链网络终端信息有效存储方法。依据终端需求得到相关知识及与知识相应的原子知识集合,优化原子知识集合,得到清洗时需要的原子知识序列,也就是数据清洗顺序,依据清洗序列实现多数据链网络终端信息清洗。将信息清洗结果代入信息存储中,利用Takens嵌入定理,得到信息流相空间重构过程中的光滑流场函数,以此计算多数据链网络终端信息时间序列嵌入矢量概率密度的特征。依据网络终端信息流相空间重构,得到多径重组规律,采用所得规律给出信息关联维特征集合。利用模糊C均值聚类实现关联维特征的聚类,引入信息存储区域自适应识别模型,完成多数据链网络终端信息存储。仿真结果表明,所提方法信息存储平均耗时为11s,存储的空间占用较少。上述方法具有较强的可行性,可实现多数据链网络终端信息有效存储。

In this paper,we presented an effective storage method for terminal information of multi-data link networks based on correlation dimension clustering.According to terminal requirements,we obtained the relevant knowledge and the atomic knowledge set corresponding to the knowledge.Through the optimization of atomic knowledge set,we obtained the atomic knowledge sequence required by clean,namely the data cleaning sequence.Based on the cleaning sequence,we realized the cleaning of multi-data link network terminal information.Then,we took information cleaning result into the information storage,and used Takens embedding theorem to get smooth flow field function during the reconstruction of information flow phase space.On this basis,we calculated the probability density feature that vector was embedded in the multi-data link network terminal information time series.According to the phase space reconstruction of network terminal information flow,we obtained the multi-path recombination law.After that,we used this law to get the set of information correlation dimension features.In addition,we used the fuzzy C-means clustering to cluster correlation dimension features.Finally,we introduced the adaptive recognition model of information storage area to complete the storage for network terminal information in multi-data link.Simulation results show that the average information storage time of the proposed method is 11 s,which needs the less storage space.Meanwhile,the method has strong feasibility,which can effectively store the terminal information of multiple data link network.