Anomaly process detection using negative selection algorithm and classification techniques

Artificial immune system is derived from the biological immune system. This system is an important method for generating detectors that include self-adaption, self- regulation and self-learning which have self/non-self-detection features. This method is used in anomaly process detection where the anomaly is non-self in the system. We present a new combining technique for anomaly process detection. This combined technique is a unification of both negative selection and classification algorithm. The main aim of the proposed techniques is to increase the accuracy in this system while decreasing its training time. In this research, CICIDS 2017 and NSL-KDD dataset with different sets of features and the same number of detectors are used. This paper presents a framework for detecting anomaly processes on a host base computer system which is established on the artificial immune system. We evaluate our technique using machine learning algorithms such as: logistic regression, random forest, decision tree and K-neighbors. Moreover, we use WEKA tool classification to perform a correlation based feature selection on the dataset.