Hybrid clustering method based on data safety

Abstract

The invention is suitable for the data mining technology field and a hybrid clustering method based on data safety. The method comprises steps that a rotation disturbance method in combination with linear transformation is utilized to carry out disturbance to an initial matrix; a density peak clustering algorithm is utilized to construct an initial mass center solution space; a gravitational search algorithm is utilized to carry out iteration to complete data point clustering. The method is advantaged in that the rotation disturbance method in combination with linear transformation is utilized to carry out disturbance to the initial matrix, a similarity measure function of clustering is guaranteed to be unchanged, data privacy processing on original data is further carried out, the density peak clustering algorithm is then utilized to select a mass center according to data point density characteristics to construct the initial mass center solution space, lastly, the gravitational search algorithm is utilized to carry out clustering of the data after disturbance according to the similarity measure function till clustering of all the data objects is accomplished, a total mean square quantization error of clustering quality evaluation is smallest, optimal data clustering is realized, data effectiveness and data privacy protection are guaranteed.

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