Blind regression modeling method for protecting data privacy in mobile group perception, and updating method

Abstract

The invention provides a blind regression modeling method for protecting data privacy in a mobile group perception system. Blind regression modeling is achieved by the interaction between a mobile perception node and a mobile perception server and can be summarized as the following steps: selecting a ''clean'' perception data subset, constructing a rough global model and performing global regression model refinement. The invention further provides an updating method of a blind regression model built by the blind regression modeling method. Model updating performed by using new perception datacan be summarized as the following steps: constructing a new rough global model and performing new regression model refinement. According to the method provided by the invention, an aggregation resultis exchanged between the mobile perception server and the mobile perception node to ensure that the contents of the perception data are not disclosed; and the incremental model updating is adopted toreduce the communication and calculation costs of the mobile perception node. By adoption of the method provided by the invention, the effects of protecting the perception data privacy, weakening theinfluence of abnormal data on the regression model, improving the model accuracy and realizing lightweight model updating are achieved.

Description

technical field [0001] The invention relates to a blind regression modeling and updating method for protecting the privacy of sensing data in a mobile group sensing system, in particular to a method for identifying "Clean" perception data subsets, and gradually refine and update the regression model. Background technique [0002] In recent years, personal intelligent mobile terminals (such as smart phones, tablet computers, etc.) have achieved rapid development in terms of processing power, embedded sensor performance, storage capacity, and wireless data transmission rate. The continuous enhancement of perception ability, coupled with its huge number of possessions, has created a new way to realize large-scale perception, that is, mobile group perception. The core idea is to make ordinary people in daily life become the main body of perception of themselves and their surrounding environment. A typical mobile group sensing system consists of mobile sensing nodes, mobile sens...

AI Technical Summary

Problems solved by technology

2. As the basic unit of perception, untrained ordinary mobile device users will inevitably have quality problems such as inaccurate, incomplete, and inconsistent perception data

As a result, perception data is often characterized by low data quality and a large proportion of outliers

3. The distribution of perceptual data may change over time, which means that the regression model needs to be constantly updated

4. The computing, communication capabilities and power of mobile devices are limited. If regression modeling imposes heavy computing and communication tasks on them, it will weaken the willingness of mobile nodes to participate in sensing tasks

It can be seen that in mobile group sensing, there are problems such as sensing data involving privacy, outliers affecting modeling accuracy, time-varying sensing data requiring model update, and sensory node resource limitations, which make accurate regression in mobile group sensing system Modeling is very difficult

[0005] In the existing perception data analysis, privacy protection technologies mainly include three categories: 1. Based on homomorphic encryption and other methods, detect the distance between data points to identify outliers as outliers. However, "leverage points" in regression model estimation may be Misjudgment; 2. Random scrambling methods that add random noise to sensory data, which will cause data distortion and affect model accuracy; 3. Least squares regression methods based solely on matrix block technology. method is very sensitive to outliers, which can lead to invalid estimates

Therefore, traditional privacy-preserving data analysis techniques are ineffective in regression modeling of mobile crowd-aware data

Images