Frequent item set mining method for local differential privacy protection based on singular value decomposition

Abstract

The invention discloses a frequent item set mining method for local differential privacy protection based on singular value decomposition. The frequent item set mining method specifically comprises the following steps of: step 1, frequency of frequent items is estimated; step 2, a server side establishes an initial matrix by using the result of the frequency estimation of the frequent items in the step 1, performs singular value decomposition on the matrix to obtain a left matrix and a right matrix, and sends the left matrix and the right matrix to a user side; 3, the user side establishes a matrix according to locally owned items, calculates a corresponding singular value matrix according to the received left and right matrixes and the frequent item set group, and uploads the interference of the singular value matrix to a server; 4, a server side carries out aggregation analysis on the received singular matrixes, mines out the most frequent item set records, updates the initial matrix, calculates new left and right singular matrixes and issues the new left and right singular matrixes to the user side; and 5, the steps are repeated 3-4 for K times, and a final top-K frequent item set is mined by the server side. The mining result is accurate, and the item set information privacy of the user is strictly protected.

Description

technical field [0001] The invention belongs to the technical field of information security. Background technique [0002] With the development of social economy and science and technology, the rapid development and improvement of information technology, including computer technology, communication technology, sensor technology, etc., has promoted the development of people's modern family life in a more convenient and comfortable direction. The term smart home has also moved from concept to reality and has become a familiar vocabulary. The intelligence of smart home is mainly reflected in the automatic execution process of related smart devices. Usually, there is a trigger condition for sensor perception. When the condition is met, the system will designate the corresponding device to perform corresponding actions. Such a trigger-action combination is the cornerstone of intelligent realization in smart home. The system needs to set a reasonable combination to help users qui...

AI Technical Summary

Problems solved by technology

However, in this method, users need to upload multiple data, which not only divides the privacy budget and leads to increased noise, but also increases communication overhead.

Images