프로세스마이닝 정리

#role of a token

  1. physical object - product, part, drug
  2. information object - message, signal
  3. collection - warehouse with parts

#role of a place

  1. communication medium - telephone line, middleman
  2. buffer -queue or post bin
  3. state condition

transition

  1. event - starting an operation, a change seasons
  2. transformation of an object - updating a database
  3. transport

network structures

  1. Causality
  2. Parallelism(AND-split,AND-join)
  3. Choice(XOR-split,XOR-join)
  4. Iteration(XOR-join,XOR-split)

#getting the data

  1. process model - event logs
    • discovery
    • conformance
    • enhancement

log

  1. process consists of cases
  2. a case consists of events
  3. events are ordered
  4. events can have attributes
  5. attribute names are activity,time,costs,and resource

XES(eXtensible Event Stream)

  1. predecessor : MXML
  2. tools such as ProM

XES Schema

  1. event -> Trace -> Log

challenges when extracting event logs

  1. correlation
  2. timestams
  3. snapshots
  4. scoping

process discovery = play-in

  1. play in -> event log => process model
  2. play out -> process modle => event log
  3. replay -> event log => process model (predictions,recommendations)

challenges - evaluating the discovered process

  1. Fitness
  2. Precision -> avoid underfitting
  3. generalization -> avoid overfitting
  4. simplicity

relations

  1. causality -> x->y iff x>y and not y>x
  2. parallel -> x   y iff x>y and y>x
  3. choice -> x#y iff not x>y and not y>x

#a-algorithm prtri-net 도출 알고리즘

  1. TI -> 시작 transition
  2. TO -> 종류

프로세스 마이닝 정의

  1. 프로세스를 도출하고, 모니터링하며, 개선하기 위해 이벤트 로그로부터 유용한 지식을 추출하는 기법입니다

BPM라이프사이클 7단계

  1. 구현 단계를 제외한 모든 단계에 활용할 수 있다

  2. 재설계 -> 분석 -> 구현 ->재구성 -> 실행 -> 조정 -> 진단 -> 재설계

프로세스 마이닝 프로젝트 수행 5단계

  1. 계획 및 타당성 평가
  2. 추출
  3. 프로세스 흐름 모델 도출
  4. 이벤트 로그 연결
  5. 통합 프로세스 모델 생성
  6. 운영 지원