An analysis and research on wireless network security dataset

摘要

Intrusion detection of wireless network has become a hot topic of research. Aegean WiFi Intrusion Dataset (AWID) is the most commonly used and open one among the intrusion detection dataset. However, AWID is characterized by character data and unbalance between attack data and normal data, which might affect the evaluation of the Intrusion Detection System (IDS). Different researchers deal with AWID in different ways, which leads to the fact that IDS research work based on this dataset cannot be evaluated by a set of unified standards. Therefore, the HWKS(Hex-Word2Vec- KMeans-Smote) data processing method is proposed by this paper, and an improved version of AWID–N-AWID(New-AWID) is proposed, in which the characteristic of character data and data imbalance is eliminated. The experimental results show that the HWKS is worked and demonstrated, and the N-AWID is more effective and challenging.