[Chapter 11] Process Monitoring

Question 1

How many ways there are for approaching process monitoring?

Answer 1

2 ways: performance dashboards and process mining.

Question 2

How many types of performance dashboards there are? Explain.

Answer 2

• Operational dashboards (runtime)
  Aimed at process workers and operational managers.
  Receives the data of current executing cases.
  WIP, resource load.
• Tactical dashboards (historical)
  Aimed at process owners/managers.
  Concern on the overall process performance on an aggregated level, not look at singular cases but in all batch of cases running in a certain period of time.
  CT, resource utilization.
• Strategic dashboards (historical).
  Aggregate and keep performances that are relevant for the overall success of the company (like customer satisfaction, customer complaint, customer feedback and connection less than agreed time).
  Customer satisfaction/feedback.

Question 3

How many levels process mining has?

Answer 3

4 levels.
- Discovery: use event-log to make a business process.
- Conformance checking: use event-log and model and see if the behavior is expressed in it.
- Deviance: use 2 event-logs and see the differences.
- Performance: use event-log and model and enhance the model.

Question 4

What is an event-log? Can they be represented as a graph? Also, how is the repetition rate calculated?

Answer 4

Event logs are often a table. Need to have column for activity name, timestamp, event id, resource, and cost.
Yes, they can be.
Repetition rate r = 1 - (sum(time)/n_executions)/(sum(time)/n_cases)

Question 5

Explain the alpha miner business process approach.

Answer 5

In the notes.

Question 6

Explain the heuristic miner business process approach.

Answer 6

In the notes.

Question 7

Explain what is conformance checking.

Answer 7

Identify such parts of sequences in the event-log that are not in line to the process model and observe deviations.

Question 8

Which are the 5 steps for process mining in practice?

Answer 8

I. Frame and plan the problem.
II. Collect the data.
III. Analyze: look for patterns.
IV. Interpret and create insights.
V. Create business impact.

Question 9

Explain the steps for split miner.

Answer 9

1) Construct the graph.
2) Remove the parallel relations.
3) Filter by some value t threshold.
4) Splits discovery.
5) Joins discovery.