IT
IT
IT
Set of flashcards Details
| Flashcards | 25 |
|---|---|
| Language | Deutsch |
| Category | Micro-Economics |
| Level | University |
| Created / Updated | 12.01.2023 / 13.01.2023 |
| Weblink |
https://card2brain.ch/box/20230112_it
|
| Embed |
<iframe src="https://card2brain.ch/box/20230112_it/embed" width="780" height="150" scrolling="no" frameborder="0"></iframe>
|
Create or copy sets of flashcards
With an upgrade you can create or copy an unlimited number of sets and use many more additional features.
Log in to see all the cards.
Pais: Systems that manage and execute operational processes involving people, applications and/or information systems
Domain-Specific PAIS: ERP, CRM, Supply Chain Management (SCM), Product Lifecycle Management (PLM).
Domain-Agnostic PAIS: Issue Tracking systems (Jira), Document Management Systems (DMC), Business Process Management Systems (BPMS), Workflow Management
Business Process Management Systems
supports the design, analysis, execution and monitoring of business processes based on explicit process models. E.g. Comunda, Flowable, Bizagi Studio
Workflow Management - Definitions:
Worklist (Task List)
Worklist Handler (Task Manager)
Interface (API)
Workflow Engine
- List of tasks assigned to a person or role
- Management of the workflows of the workers
- A connector to another application running in a separate environment (Service Interface)
- Run-tim engine which controls the execution
Technical Process Flow
User Inclusion
Decision Support
Service Integration (Delegate Java code or Connector REST / SOAP)
Why should we deal with data?
1) What happend?
2) What will happen?
3) Why did it happen?
4) What is the best that can happen?
Process Mining as the bridge between:
Data Science and Process Science
What is Process Mining:
1) Make it possible to reconstruct and analyse business processes based od digital traces in IT systems (EVENTLOGS)
2) Make it possible to implicit and otherwise hidden process knowledge contained in data and thus make it tangible and transportable
3) is data-driven and process-centric
Event Data as the Starting Point:
Case ID
TImestamp
Activity
Attributes
Events
Instances
Process Discovery:
Discover a process model based on the event log, under the condition not be "overfitting" and "underfitting".
(Underfitting: exclude so many case variants, that the process doens't fit to realisty)
(Overfitting: Everything is considerd form the eventlogs).
Conformance Checking:
Existing process model is compared with an event log of the same process, aiming to indicate disagreement between the model and event log
Performance Checking:
reveal performance problems (e.g. unimely completion of a case, missed deadline, tardiness, as well as recurring quality problems) and annotate the process model with frequency and time information
Comparative Process Mining:
Inout are multiple event logs (e.g. from the same process perfrrmed at different locations) and aims to identify execution gaps
Predictive Process Mining:
Use of ML techniques to predict compliance and performance problems
Action-Oriented Process Mining:
Turn diagnostics of the current state of a process into improvements actions (e.g. case waiting time is too long -> assing additional resources, detect repetivive work which can be automated using Robotic Process Automation).
Perspektives; Aspects that Process Mining aims to analyse
Control-flow perspektive (order of acitivities)
Organizational perspektive (information resources hidden in the log, actors)
case perspective (properties of cases)
time perspective (timing and frequency)
decision perspective
Minimum requirements for process mining
Case ID
Activity
Timestamp
Challlenges when extracting event logs
Correlation (events need to be related)
timestamps (need to be ordered per case)
snapshots (different lifetime)
scoping (which tables to incorporate)
granularity
Three ways of relating event logs and process models:
Play-In: Example behavior (event logs) is taken as input, and the goal is to construct the model (e.g. Alpha Algorithm using PETRI NET)
Play-Out: given a model, it is possible to generalte behavior (generated logs)
Replay: Uses an event log and a process model as input
Examples time perspektive:
Avg. Time customer spend waiting for a seat
avg. time to clean the plates
avg. time until food can be served
Examples case perspektive:
Path discovery in the process
Examples organisation perspektive:
As we can relate the time perspective to the resources
Data Mining - 2nd Pillat of Process Mining (2 Approaches)
Supervised and Unsupervised.
Data Mining results may be both descriptive and predictive.
Decision trees, association rules, and regression functions say somethings about the data set used to learn the model.
They can also be used to make predictions for new instances.
Datra Mining - Supervised Learning
Response Variables - dependent
Predicotr Variables - independent
Type of response variable: Cliassification (categorical variables9 and Regression (numerical)
Datra Mining - Techniques
Decision Tree
k-Means Custering
Datra Mining - Unsupervised Learning
Clustering
Pattern discovery
-
- 1 / 25
-
