A Traceability Method for Sampling-Constrained Active Learning

Abstract

The present invention provides a sampling-limited active learning traceability method to evaluate the SLP strength under the false source mechanism in the information physics system; considering the limited system cognition of the traceability agent, which makes the data labeling ability limited, the sampling-limited The traceability method; considering that unconstrained random walk will lead to the phenomenon that the traceability period is too long, the network traffic identification model is used to optimize the traceability strategy; considering that the false traffic generated by the false source mechanism may be much higher or lower than The number of normal traffic, in order to solve this category imbalance and potential outliers, an active learning strategy combining uncertainty, sample double representativeness and diversity is adopted. This method has a better ability to capture source nodes than the random walk traceability method, and the proposed active learning strategy can balance positive and negative samples more effectively than other active learning methods, and has better sampling performance.

Description

technical field [0001] The invention relates to the technical field of active learning, in particular to a source tracing method for limited sampling active learning. Background technique [0002] At present, the existing source location privacy (SLP) research mainly uses random walk traceability, and its random method does not conform to the actual traceability model. In order to solve the problems of limited system cognition, unconstrained random walk, imbalance and potential outliers due to the traceability agent, an intelligent active learning traceability method is urgently needed. Contents of the invention [0003] In order to realize the traceability method of sampling-limited active learning, the present invention proposes a sampling-limited active-learning traceability model to evaluate the SLP strength under the false source mechanism in cyber-physical systems. This method includes the establishment of an intelligent traceability model , the network traffic iden...

AI Technical Summary

Problems solved by technology

In order to solve the problem of limited system cognition of traceability agent, unconstrained random walk, imbalance and potential outliers, an intelligent and active learning traceability method is urgently needed

Images