Multi-dimensional sequencing and searching method for mass data

Abstract

The invention discloses a multi-dimensional sequencing and searching method for mass data. The method comprises the following steps: clustering files according to field relevance of files in database to obtain search trees having similar clustering structures; clustering the clustering structures in different fields and forming similar search trees; obtaining search vectors submitted by users and presenting vectors as search supraspheres; and obtaining a suprasphere having most intersections with the search supraspheres by searching the positioning relationship of supraspheres represented by nodes of similar search trees and search supraspheres, searching in the next layer of nodes of supraspheres till nodes of leaves, searching neighboring nodes on the left and right and according to correlation proportions, returning most-related k file lists and file vectors in the nodes. The multi-dimensional sequencing and searching method for mass data has the following beneficial effects: a DIK-MEDOIDS algorithm in the environment of big data has obvious advantages; and search speed of data is increased and accuracy of data is improved.

Description

technical field [0001] The invention relates to the technical field of data search, in particular to a multi-dimensional sorting search method for massive data. Background technique [0002] The information privacy of the wireless body area network includes various physiological parameters of the user, and a large amount of private data is stored in the cloud server. The ciphertext retrieval technology is an effective method to solve the privacy security problem in the cloud environment. In the current data protection technology, encryption algorithms can better protect data, but encryption and decryption calculations will have a great impact on system efficiency; the data splitting and reloading strategy is more efficient, but it does not affect the structure and physical properties of the cloud platform. Hierarchical dependencies are too large. Therefore, finding a balance between data availability and security is the most critical issue in cloud storage platform applicat...

AI Technical Summary

Problems solved by technology

Tian Xue et al. optimized the ciphertext retrieval framework MRSE, and proposed a new ciphertext index structure: MRSE-SS, which introduced the similar query tree structure into the ciphertext index framework to improve the efficiency of multi-keyword sorting and retrieval, and proposed A dynamic clustering algorithm DK-MEDOIDS, the clustering process changes dynamically with the increase of the number of documents, which is suitable for the ciphertext retrieval scene in the cloud computing environment, but in this method, the worst time complexity when constructing a hypersphere will be up to o(n2), and if, when the query algorithm passes back documents, the number of documents in the most relevant hypersphere is less than the k queried, then this method cannot solve the problem

Images