WiFi Miner: An Online Apriori-Infrequent Based Wireless Intrusion System

Abstract

Intrusion detection in wireless networks has become a vital part in wireless network security systems with wide spread use of Wireless Local Area Networks (WLAN). Currently, almost all devices are Wi-Fi (Wireless Fidelity) capable and can access WLAN. This paper proposes an Intrusion Detection System, WiFi Miner, which applies an infrequent pattern association rule mining Apriori technique to wireless network packets captured through hardware sensors for purposes of real time detection of intrusive or anomalous packets. Contributions of the proposed system includes effectively adapting an efficient data mining association rule technique to important problem of intrusion detection in a wireless network environment using hardware sensors, providing a solution that eliminates the need for hard-to-obtain training data in this environment, providing increased intrusion detection rate and reduction of false alarms.

The proposed system, WiFi Miner solution approach is to find frequent and infrequent patterns on pre-processed wireless connection records using infrequent pattern finding Apriori algorithm proposed by this paper. The proposed Online Apriori-Infrequent algorithm improves the join and prune step of the traditional Apriori algorithm with a rule that avoids joining itemsets not likely to produce frequent itemsets as their results, there by improving efficiency and run times significantly. An anomaly score is assigned to each packet (record) based on whether the record has more frequent or infrequent patterns. Connection records with positive anomaly scores have more infrequent patterns than frequent patterns and are considered anomalous packets.