Intrusion detection systems (IDSs) must meet the security goals while minimizing risks of wrong detections. In this paper, we study the issue of building a risk-sensitive intrusion detection model. To determinate whether a system calls sequence is normal or not, we consider not only the probability of this sequence belonging to normal sequences set or intrusion sequences set, but also the risk of a false detection. We define the risk model to formulate the expected risk of an intrusion detection decision, and present risk-sensitive machine learning techniques that can produce detection model to minimize the risks of false negatives and false positives. Meanwhile, this model is a hybrid model that combines misuse intrusion detection and anomaly intrusion detection. To achieve a satisfying performance, some techniques are applied to extend this model.