**Abstract**

The Dendritic Cell Algorithm (DCA) is a bioinspired, population-based, supervised binary classifier, designed for anomaly detection in communication networks. The proposed model is inspired by the behavior of Dendritic Cells and Danger Theory. The main contribution of this research addresses two contemporary challenges of Network-based Intrusion Detection Systems, namely feature selection and generalization capabilities to improve classification performance. Feature selection improvement is achieved by using information gain and mutual information. A Decision Tree model is incorporated as a classification mechanism in order to improve accuracy, as a substitute to the classification threshold of the DCA. The proposed model is assessed using two publicly available datasets, namely UNSW-NB15 and NSL-KDD. Experimental results are compared against state of the art bioinspired and machine learning approaches for binary classification. The proposed approach provides competitive results when compared to other state of the art approaches, such as Support Vector Machines, and Artificial Neural Networks, achieving a 97.25 and 93.28% accuracy for the UNSW-NB15 and NSL-KDD datasets, respectively. Future challenges include multi-class classification, further performance improvements, and online detection.

**Keywords:** Anomaly detection, Dendritic Cell Algorithm, Decision Tree, binary classifier, Danger Theory, Artificial Immune System
