Ontology Based Log Analysis of Web Servers Using Process Mining Techniques

• Published in: 2018 10th International Conference on Electrical and Computer Engineering (ICECE) pp. 341 - 344
• Main Authors: Barua, Dipto, Rumpa, Nishat Tasnim, Hossen, Sakhawat, Ali, Muhammad Masroor
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-12-2018
• Subjects: Analytical models; Computational modeling; Data mining; Data models; Ontologies; Ontology; Process Mining; Tag Prediction; Unified modeling language; Web servers
• DOI: 10.1109/ICECE.2018.8636791

Process mining techniques provide an opportunity for discovering information about process models in organizations or web based systems based on the execution of data registered in event logs. Lack of logical inference and observable abstractions on the data of the generated process models limit our analysis to nothing more than the bare bone structure regarding processes and connections among them. Ontology is the most intuitive yet powerful solution that can connect traditional data mining techniques and process mining techniques enriching the analysis based on specific or in general queries on the event logs or databases. Our objective of this paper is to translate the entirety of the process mining techniques to ontological representations. Performance matrices and process discovery techniques are the main aspects that need to be converted to ontological concepts. For analysis on real world event logs, we have chosen stack-overflow data in order to provide a tag based prediction technique.