Process Management with Big Data and Deep Learning
Process mining is the analysis of event logs. Many information systems produce event logs that capture the actions of their users. Examples of event logs are page requests of web-servers and business-object method calls in ERP systems. Process discovery is that area of process mining that deals with the identification of the processes followed by system users, for example the process of ordering a product on an e-commerce web-site, or the process of scheduling a manufacturing order in an ERP system. Process discovery is important when the underlying systems are not process-aware. Another area of process mining is compliance assurance, determining whether the actually executed process conforms to a prescribed one.
Modern information systems, such as web-crawlers, web-servers, ERP systems, databases, etc., are increasingly distributed, with replicated instances deployed on multiple physical machines, for example as part of a load-balancing architecture or for geographic proximity to users. Given the distributed nature of event logs, it is natural to look for a distributed way to mine these for processes. The Map-Reduce approach is a scalable means of analyzing distributed data and performing distributed computations on such data. In this ongoing research project, I investigate how well-known process mining algorithms can be implemented using Map-Reduce. Such implementations take advantage of the natural fit between distributed event logs and distributed computations, to make the algorithms scalable to large data sets.
Related to this is the use of modern Deep Learning frameworks for process prediction. Predicting the future behaviour of a process at runtime is important for the active and timely management of the process to reduce operational risk and the probability of process failure. The size of event data that is available for past and present process instances requires modern Deep Learning frameworks to process them. Recent advances in neural network architectures and GPU processing capabilities make the application of Deep Learning methos feasible.
Publications on this Topic
- Mehdiyev, N., Evermann, J., and Fettke, P.: A Novel Business Process Prediction Model Using a Deep Learning Method. BISE Business and Information Systems Engineering (accepted May 22nd, 2018).
- Preprint
- Full Paper (external link)
- Mehdiyev, N., Fettke, P. and Evermann, J.: A Multi-Stage Deep Learning Approach for Business Process Event Prediction. 19th IEEE Conference on Business Informatics (CBI), Thessaloniki, Greece, July 24-27, 2017.
- Full Paper (external link)
- Evermann, J., Rehse, J.-R., and Fettke, P.: XES Tensorow - Process Prediction using the Tensorflow Deep-Learning Framework. Forum of the Conference on Advanced Information Systems Engineering (CAiSE), Essen, Germany, June 12-16, 2017.
- Preprint (PDF)
- Preprint (arXiv)
- Full Paper (external link)
- Evermann, J., Rehse, J.-R., and Fettke, P.: Predicting Process Behaviour Using Deep Learning" Decision Support Systems (forthcoming) (accepted April 5th, 2017)
- BibTeX
- Preprint (PDF)
- Preprint (arXiv)
- Full Paper (external link)
- Evermann, J., Rehse, J.-R., and Fettke, P.: Process Discovery from Event Stream Data in the Cloud - A Scalable, Distributed Implementation of the Flexible Heuristics Miner on the Amazon Kinesis Cloud Infrastructure. CloudBPM Workshop on Business Process Monitoring and Performance Analysis in the Cloud at the 8th IEEE International Conference on Cloud Computing Technologies and Science (CloudCom 2016) .
- BibTeX
- Preprint (PDF)
- Full Paper (External Link)
- Evermann, J., Rehse, J.-R., and Fettke, P.: A Deep Learning Approach for Predicting Process Behaviour at Runtime. PRAISE International Workshop on Runtime Analysis of Process-Aware Information Systems at the 14th International Conference on Business Process Management (BPM). (accepted July 4, 2016).
- BibTeX
- Preprint (PDF)
- Evermann, J., Thaler, T. and Fettke, P.: Clustering Traces Using Sequence Alignment Business Process Intelligence Workshop at BPM 2015 (accepted June 26th, 2015).
- BibTeX
- Preprint (PDF)
- Full Paper (External Link)
- Evermann, J.: Scalable Process Discovery using Map-Reduce. IEEE Transactions on Services Computing, 9 (3), 469-481.
- BibTeX
- Preprint (PDF)
- Full Paper (External Link)
- Evermann, J. and Assadipour, G. (2014) Big Data meets Process Mining: Implementing the Alpha Algorithm with Map-Reduce. ACM Symposium on Applied Computing, March 24-28, Gyeongju, Korea.
- BibTeX
- Full Paper (External Link)