Oppositional tunicate fuzzy C‐means algorithm and logistic regression for intrusion detection on cloud

Summary Cloud computing is the most popular emerging computational model in information technology. The data communication security requirement is increased day by day with the development of cloud computing. Virtualization facilitates cloud computing architecture, which raises a few security concerns. In the previous year, many new research approaches were introduced, resulting in more sophisticated and time‐consuming results, as well as less flexibility, a lower attack detection rate, and so on. In this article, we proposed a logistic regression‐based oppositional tunicate fuzzy C‐mean (LR‐OTSFCM) methodology for the detection of cloud intrusion. The major contribution of this article is attack detection from the cloud environment. The proposed work is classified into four stages, namely, preprocessing, partitioning, cluster handling, and intrusion detection stages. In the first stage, we use the text data as the input; then, it is preprocessed. Second, the feature selections are performed via a logistic regression model. The combination of the oppositional tunicate swarm algorithm (OPTSA) and fuzzy C‐mean clustering (FCM) model is to perform data clustering. Third, the cluster handling stage is handled by cluster expansion and integration. Fourth, the clustering result builds a profile of normal and abnormal behavior. The local optimum avoidance ability of OPTSA is tested using four benchmark problems, namely, composition functions, unimodal, multimodal, and fixed dimensions of multimodal functions with 20 benchmark functions. Ultimately, the proposed LR‐OTSFCM method is compared with state‐of‐art approaches such as ANN, LR‐HID, ML‐IDS, and En‐ABC. The proposed scheme effectiveness is carried out by using four datasets such as, namely, synthetic, NIB, CICID 2017, and CIDD datasets. The proposed method demonstrated a superior attack detection rate.