PhD Research Seminar: A Hybrid Process Mining and Machine Learning Approach to Detection and Prediction of Anomalies for Storage Area Networks
Nowadays storage area networks need failure-prediction systems to improve fault tolerance. This is why process mining and machine learning techniques have recently received a significant amount of attention by storage architectors and ML researchers. Indeed, these techniques allow both an automatic process discovery and preventive analytics for anomalies detection and prediction. The global aim of this research is to shorten periods of storage area network downtime. The experimental results obtained by different network architectures for clustering of logs and constructing process models in pm4py library compared with analogs will be considered and analyzed.
Карпов Максим Евгеньевич
Научно-учебная лаборатория методов анализа больших данных: Стажер-исследователь