An Efficient Intrusion Detection Method Based on Federated Transfer Learning and an Extreme Learning Machine with Privacy Preservation

Current network security is becoming increasingly important, and intrusion detection an effective method to protect the from malicious attacks. This study proposes algorithm FLTrELM based on federated transfer learning extreme machine improve effect of detection, which implements data aggregation through facilitates construction personalized for all organizations. first builds a model solve problem insufficient samples probability adaptation, then uses learn privacy without sharing training under mechanism, finally obtains model. Experiments NSL-KDD, KDD99, ISCX2012 datasets verify that proposed can achieve better results robust performance, especially small new intrusions, protects privacy.