A Harmony Search-Based Feature Selection Technique for Cloud Intrusion Detection

Recently cloud computing has enjoyed widespread patronage due to its economy of scale and flexibility. However Cloud computing is confronted with security challenges. Intrusion detection can be used to protect computer resources from unauthorized access. However the presence of insignificant features in Intrusion Detection dataset may have a negative effect on the accuracy of Intrusion Detection System (IDS). Feature selection is utilized to remove noisy and insignificant attribute to improve IDS performance. However, existing feature selection techniques proposed for cloud IDS cannot guarantee optimal performance. Therefore, this research article proposes a Harmony Search based feature selection technique to improve the performance of cloud IDS. The attributes selected were assessed using Random Forest classifier and experimental results of the Harmony Search based technique achieved an attack detection rate of 79% and a false alarm rate of 0.012%. In addition performance comparison shows that the proposed Harmony search outperforms existing feature selection technique proposed for cloud IDS.