RT @wvdaalst: Breaking news for anyone interested in #ObjectCentricProcessMining! Today, we released OCEL 2.0. Visit .

Some updates on my personal website including major additions (Korea & Australia) to the Schnitzel world map!.

8/8 💻: Our paper also includes a comprehensive taxonomy for process concept drifts. Plus, the code is publicly available for you to explore, extend, and adapt with new techniques or evaluation goals: #OpenScience #VLDB2023.

7/8 💪 Robustness: Finally, we evaluate algorithm resilience to noise, using two noise definitions (adding unfitting behavior & removing fitting behavior). Find out which algorithms thrive, and which falter, under noise.

6/8 🔄 Parameter Sensitivity: Our research assesses performance across an algorithm's parameter space, identifying those requiring fine-tuning vs. those with broader adaptability.

5/8 🚀 Scalability: We highlight the average computation time per event, revealing significant differences that impact practical application.

4/8 🎯 Versatility: Our evaluation tests algorithm performance on 16 distinct business process change patterns, revealing unique strengths and weaknesses.

3/8 ⏱️ Latency: We analyze the delay between detected drifts and ground-truth concept drifts for each algorithm.

2/8 🔍 Accuracy: We measure algorithm accuracy using true/false positives from bipartite matching between detected & ground truth drifts, plus a sensitivity analysis to the bipartite matching.

1/8 📊 Curious about how process concept drift detection algorithms stack up? #VLDB2023 features our in-depth experimental evaluation of 8 algorithms across 6 evaluation goals. Learn more: with Cameron Pitsch, Tobias Brockhoff, and @wvdaalst

RT @Celonis: X-Ray vs. MRI. The old way vs. The new way. The future of process mining. The future of business performance. #ProcessMin….

More than two years ago, I joined @pads_rwth and presented a tweetish description of my master's thesis on our trip to @dagstuhl. Now, the corresponding paper was accepted at #InformationSystems as my first journal paper🥳Read the full text here:.@wvdaalst

RT @gnpark7: At @icsoc2022, @jn_adams presents our co-authored paper: “A Framework for Extracting and Encoding Features from Object-Centri….

I will be presenting our paper on Thursday, 14:30. If you are at ICSOC feel free to attend!.Link to paper:.@gnpark7 @cortado_tool @wvdaalst.

We finally reached beautiful Sevilla for this year's @icsoc2022! Looking forward to many interesting presentations. From @pads_rwth @gnpark7, @cortado_tool, Tsung-Hao Huang, and me will be presenting over the next days.

RT @wvdaalst: Interested to see how #ProcessSphere implements #ObjectCentricProcessMining? Watch the #Celosphere opening keynote. This majo….

Today, @Celonis announced the move of their platform to object-centric process mining. This will not only provide Celonis' (Germany's most valuable startup at 13.2b$) customers with value but it is also an important signal for the symbiosis of research and industry in Germany.

All these contributions are now bundled in the #Python library ocpa (published in Software Impacts, GitHub: .

Starting from object-centric event logs, we.translated the case concept from event sequences to graphs, provided variant visualization, checked model conformance, analyzed cross-process performance, and provided feature extraction for machine learning on multiple processes.

It is a great feeling when research quickly becomes adopted in industry. Two years ago we started a journey with one question: How to apply #processmining to multiple, interacting processes?. @wvdaalst @gnpark7 @cortado_tool @Alessan75384183 . ↓ Read more