Temporal Analysis for Process Mining

This codebase can be split into two parts: feature computation and decision tree classification for temporal root cause analysis in conformance checking.

General settings:

• the code was developed for Python 3.5

Feature Computation:

• parameters and other settings (such as working-directory) must be set in the feature_config.ini in advanced:
  • base_path: working-directory
  • log_file: path to XES event log
  • report_file: path to conformance report (CSV)
  • event_sets: path to event sets (CSV)
  • case_limit: number of cases to import (set -1 to import all)
  • number_jobs: number of parallel jobs
  • plots: True or False
  • feature_level: process granularity(ies) e.g. event,set,case (NO spaces!)
• the conformance report must only contain the case information. If necessary, leading and subsequent rows have to be deleted in advance.
• event sets have to be stored in a .csv file. Events of one set have to be separated by |
• use the compute_features.py script to start the feature computation

Classification:

• use the classify.py script to start the classification task