Ontology Based Anomaly Detection for File Integration

Abstract

File integration systems enable file transfers between different systems in order to automate routine business processes. Therefore, the standardization in data exchange between different organizations or decentralized subsidiaries of an organization is achieved. However, abnormal situations may occur during the file integration process. In order to protect the persistence of integration channels, the abnormal files must be detected. For this purpose, anomaly detection is used to trace integrations continuously and to detect abnormal files instantly. In this study, an ontology based anomaly detection approach is proposed in order to detect abnormal situations in real time file integration systems. Thus, a file integration that is achieved on an electronic system will be traced and information will be given to the system administrator if any abnormalities occur during the integration process. Therefore, an abnormal situation that can stop the current file flow on file integration systems will be detected.