A Hybrid Clustering Method Based on Data Safety

Abstract

The present invention is applicable to the technical field of data mining, and provides a hybrid clustering method based on data security, which includes the following steps: using a rotation perturbation method combined with linear transformation to perturb the initial matrix; using a density peak clustering algorithm to construct an initial matrix Centroid solution space; use the gravitational search algorithm to iteratively complete the clustering of data points. In the embodiment of the present invention, the rotation perturbation method combined with linear transformation is used to perturb the initial matrix to ensure that the similarity measurement function of the clustering remains unchanged. At the same time, the original data is processed for data privacy, and then the density peak clustering algorithm is used according to the data points The method of selecting the centroid of the density characteristic constructs the initial centroid solution space, and finally uses the gravitational search method to cluster the disturbed data according to the similarity measurement function until all data objects are clustered, and the total mean square quantization error of the clustering quality is evaluated The smallest, so as to achieve the optimal clustering of data, ensure data utility and hide data privacy.

Description

technical field [0001] The invention belongs to the technical fields of data security, privacy protection and data mining, and in particular relates to a hybrid clustering method based on data security. Background technique [0002] Based on the rapid development of the Internet, complex networks have become an important channel and carrier for social relationship maintenance and information dissemination in human society. While the number of users is growing day by day, the amount of data is also skyrocketing exponentially. At this time, data mining technology has played an important role. It is very important to analyze the topological structure of complex network data, understand network functions, discover hidden laws in data, and mine hidden related information through mining data generated by Internet-based service processes. theoretical significance. [0003] Based on the powerful ability of data mining to extract hidden information from massive data, security issues...

AI Technical Summary

Problems solved by technology

[0005] The embodiment of the present invention provides a hybrid clustering method based on data security, which aims to solve the problem that existing data protection methods cannot balance data utility and privacy protection, and usually sacrifice privacy protection to improve data utility, resulting in data privacy. The problem of low protection

Images