Misuse activity in computer networks constantly creates new challenges and difficulties to ensure data confidentiality, integrity, and availability. The capability to identify and quickly stop the attacks is essential, as the undetected and successful attack may cause losses of critical resources. The anomaly-based intrusion detection system (IDS) is a valuable security tool that is capable of detecting new, previously unseen attacks. Anomaly-based IDS sends an alarm when it detects an event that deviates from the behavior characterized as normal. This paper analyses the use of the histogram-based outlier score (HBOS) to detect anomalies in the computer network. Experimental results of different histogram creation methods and the influence of the number of bins on the performance of anomaly detection are presented. Experiments were conducted using an NSL-KDD dataset.