There are many studies done to detect anomalies based on logs. Current approaches are mainly divided into three categories: supervised learning methods, unsupervised learning methods, and deep learning methods. Many supervised learning methods are used for log-based anomaly detection.