Research On The Internal Characteristics And Compression Algorithm Of Lattice Structure Data

Data cubes are the core data model of Data warehousing and OLAP,while Concept lattices are a kind of important data model in formal concept analysis theory.Data cubes and Concept lattices are widely used in the field of data analysis.This paper demonstrates the relationship between data cubes and concept lattices from two aspects: generation mechanism and structural characteristics.The experimental results show that the data cube has similarity in terms of degree distribution,clustering coefficient,average shortest path and so on.At present,data cube compression technology is a focus of data cube computing research,especially the data cube semantic compression technology has become a hot topic.BUS algorithm is an effective compression algorithm of graph,based on the structure characteristics of the lattice structure data,this paper applies the BUS algorithm to the lattice structure.At the same time,the operation time,error rate and the compression effect on the condition entropy are compared and analyzed.The primary contents are as follows:(1)Lattice data are seen as graph data,this paper demonstrates the relationship between the Data cubes and the Concept lattices in generative mechanism and structural characteristics(such as degree distribution,clustering coefficient,average shortest path and so on)through the method of theoretical analysis and experimental demonstration.(2)Based on the degree distribution rule of Data cubes structure in statistical characteristics of graph structure,a partition method combining edge partition and point partition based on distributed computing environment are proposed.(3)Based on the bottom-up graph aggregation algorithm,the lattice structure data are compressed and the compression results under different parameters conditions are analyzed.