AbstractComputer network is playing a significantly important role in our society, including commerce, communication, consumption and entertainment. Therefore, network security has become increasingly important. Intrusion detection systems have received considerable attention, which not only can detect known attacks or intrusions, but also can detect unknown attacks. Among the various methods applied to intrusion detection, logistic regression is the most widely used, which can achieve good performances and have good interpretability at the same time. However, intrusion detection systems usually confront with data of large scale and high dimension. How to reduce the dimension and improve the data quality is significant to improve the detection performances. Therefore, in this paper, we propose an effective intrusion detection model based on pls-logistic regression with feature augmentation. More specifically, the feature augmentation technique is implemented on the original features with goal of obtaining high-qualified training data; and then, pls-logistic regression is applied on the newly transformed data to perform dimension reduction and detection model building. The NSL-KDD dataset is used to evaluate the proposed method, and the empirical results show that our proposed method can achieve good performances in terms of accuracy, detection rate and false alarm rate.