作者: Satoru Koda , Yusuke Kambara , Takanori Oikawa , Kazuyoshi Furukawa , Yuki Unno
DOI: 10.1109/COMPSAC48688.2020.00-42
关键词:
摘要: This paper presents an anomalous IP address detection algorithm for network traffic logs. It is based on word embedding techniques derived from natural language processing to extract the representative features of addresses. However, extracted vanilla embeddings are not always compatible with machine learning-based anomaly algorithms. Therefore, we developed that enables extraction more addresses than conventional methods. The proposed optimizes objective functions embedding-based feature and detection, simultaneously. According experimental results, outperformed approaches; it improved performance 0.876 0.990 in area under curve criterion a task detecting attackers