Graph data compression method based on Boolean matrix decomposition

Abstract

The invention discloses a Boolean matrix decomposition-based graph data compression method, which specifically comprises the following steps of: splitting original graph data, expressing by using a self-center network, and reordering; expressing the sorted self-center network set by using an adjacent matrix to generate a sampling matrix; decomposing the sampling matrix into a product of a dictionary matrix and a sparse code matrix; performing Boolean atom mining; the atoms are linearly combined to obtain a sampling recovery matrix, and then according to a sampling node set obtained in the sampling stage, the edge connection relation between the nodes is recovered according to the sampling recovery matrix to obtain recovered graph data. According to the method, the dictionary matrix and the sparse code matrix which are obtained through matrix decomposition are constrained in a Boolean matrix decomposition mode, so that the dictionary matrix and the sparse code matrix are both Boolean matrixes, the error rate of graph data representation can be reduced, meanwhile, the accuracy of atoms obtained through representation is improved, and compression of graph data is achieved.

Description

technical field [0001] The invention belongs to the field of Internet technology, and in particular relates to a graph data compression method. Background technique [0002] In recent years, with the development of the Internet, the amount of data has also exploded. Network graphs are constructed and analyzed through information generated in social media, which provides strong support for scientific research, such as protein network analysis of protein properties, user commodities The network is used for product recommendation, the social network graph is used for user friend recommendation, etc. However, due to the limitation of memory and the continuous expansion of graph scale, many graphs can no longer be completely put into memory, which brings huge challenges to their storage and analysis. Large-scale graph data also has the problem of complex structure and high coupling, which leads to a huge time overhead when performing some information query operations, and the ca...

AI Technical Summary

Problems solved by technology

First of all, for floating-point numbers, it is difficult to judge its physical meaning. For example, for 0.5 in the dictionary matrix, it is not appropriate to judge whether it has even edges; secondly, for negative numbers that appear in the matrix, such as -1.5, it should not exist. , is forced to appear only for the error of the fitting matrix, in fact we do not want to see it

The network sparse representation will generate atoms through the dictionary matrix, and negative numbers and floating-point numbers cannot correctly judge whether they exist in the real graph structure data

Due to the above problems, the sparse representation of the network cannot be accurately represented

Images