An efficient and trusted packet anomaly detection method based on automaton

By combining the methods of neural networks and finite automata, the problem of unreliable detection results of neural network models is solved, efficient and reliable message anomaly detection is achieved, which adapts to different network environments and improves the reliability and efficiency of detection.

CN119382939BActive Publication Date: 2025-10-17ANHUI UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202411395229.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-10-08
Publication Date
2025-10-17
Estimated Expiration
2044-10-08

AI Technical Summary

Technical Problem

Existing neural network-based message anomaly detection methods lack credibility, and the internal judgment logic of the neural network model is opaque, resulting in unreliable detection results.

Method used

Combining the neural network model and finite automaton, the output results of the neural network model are verified through data preprocessing, feature selection and the migration function rules of the message finite automaton, and an automaton for detecting anomalies of message content and historical messages is constructed to improve detection reliability.

Benefits of technology

Effectively verify the output results of the neural network model, reduce false positives and missed negatives, improve detection efficiency, adapt to different network traffic and application scenarios, and enhance network security.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The application discloses a high-efficiency and credible message exception detection method based on an automaton, which comprises the following steps: firstly, pre-processing message data to obtain a message data set; then, using a random forest algorithm to calculate the importance of each message feature in the message data set, selecting message features with high importance in the message data set to form a feature data set; then, using a neural network model to train and detect the feature data set to obtain a message exception detection result; finally, inputting the output of the neural network model into a message finite automaton, and performing exception detection on the message data according to a transition function rule of the message finite automaton to verify the output result of the neural network model. The application combines the advantages of the neural network model and the finite automaton, and effectively improves the reliability and efficiency of message data detection.
Need to check novelty before this filing date? Find Prior Art