Welcome to ECS Event Abstraction (ECSEA)’s documentation!

Note

This project is under active development.

• Install
• Training
• Application
• API
• CLI

Library for Event Abstraction with event logs from Enterprise Collaboration Systems (ECS). The basic idea is that existing (e.g., observed) high-level traces are used together with the related low-level (extracted from the ECS) traces to train a model that can then be applied to other low-level traces to convert them into high-level traces.

Thus, this library provides two main functions:

1. Training: Train a model by comparing a low-level event log and a high-level (observed) event log.

2. Application: Use the created model to convert a low-level event log to a high-level event log.

Cite

Blatt, J., Delfmann, P., Schubert, P., & Just, M. (2026). Event abstraction for social process mining in Enterprise Collaboration Systems: A supervised machine learning approach for converting low-level event logs into high-level event logs. Information Systems, 139, 102703. https://doi.org/10.1016/j.is.2026.102703