マルチソースアクセスログ分析によるボット検知手法の提案

抄録

In this research, we proposed a data cleansing method to support the management and utilization of access logs in the operation of websites. We focused on detecting and removing bots hidden in access logs. The method we propose discovers the hidden features of bots by extracting access logs using multiple methods. We propose a new method that combines the hidden features of bots with rule-based methods and traditional anomaly detection methods. The proposed method can automate the same degree of accuracy as manual bot discrimination. In addition, the proposed method improves the prediction accuracy of recommender systems.