Approach for Enhancing Incident Management Process Assessment Through a Process Mining Based Approach

Description

This program is a proof of concept for an automated approach to perform Incident Management (IM) process accessment leveraging Trace Alignment. In particular, this program performs the analysis of fitness as it derives by alignment, cost analysis describing the errors causing deviation from the target model, detailed assessment of the specific events causing errors, and incident analysis basing on their type (i.e., incident categories and affected services).

Installation requirement

The following libraries are required for the correct execution:

  • pip install pm4py
  • pip install matplotlib
  • pip install numpy
  • pip install pandas

Configuration

If you want to just reproduce a simplified version of the proof-of-concept, follow the installation instructions.

If you want customize your assessment, the following configurations are settable in the file conf.py:

  • input IM log file: put the file (in format csv) in the folder named "data" and set the filename in the conf.py (fileLog parameter)
  • input target model file: put the file (in format .pnml) in the folder named "data" and set the filename in the conf.py (fileModel parameter)
  • set the parameters of the cost model with the following notation: N=detection,A=activation,W=awaiting,D=double-check,F=notification,R=resolution,C=closure
  • set the weight of missing, repetition and mismatch errors, and the severity thresholds

Installation

Download this github repositiory, set the configuration (if any) and run the command from your terminal:

$ python -u "<path_to_this_folder>\main.py"