IEEE XES Standard

Why do we need the XES Standard?

The  goal  of  the  eXtensible  Event  Stream (XES)  Standard  is  to  standardize  a  language  to  transport,  store,  and  exchange  (possibly  large  volumes  of)  event  data  (e.g., for process mining).

The spectacular growth of the digital universe,  summarized  by  the  overhyped  term  “Big  Data,”  makes  it  possible  to  record, derive, and analyze events. Events may take place inside a machine (e.g., an X-ray  machine,  an  ATM,  or  baggage  handling  system),  inside  an  enterprise  information  system  (e.g.,  an  order  placed  by  a  customer  or  the  submission  of  a  tax  declaration),  inside  a  hospital  (e.g.,  the  analysis  of  a  blood  sample),  inside a social network (e.g., exchanging e-mails  or  twitter  messages),  inside  a transportation  system  (e.g.,  checking  in,  buying a ticket, or passing through a toll booth),  etc.  Events  may  be  “life  events,”  “machine events,” or “organization events.” The term Internet of Events (IoE),  refers to all event data available. The IoE is composed of:

  • The  Internet  of  Content  (IoC):  all  information  created  by  humans  to  increase  knowledge  on  particular  subjects. The IoC includes traditional web  pages,  articles,  encyclopedia  like  Wikipedia,  YouTube,  e-books,  news-feeds, etc.
  • The Internet of People (IoP): all data related  to  social  interaction.  The  IoP  includes  e-mail,  Facebook,  Twitter,  forums, LinkedIn, etc.
  • The  Internet  of  Things  (IoT):  all  physical  objects  connected  to  the  network.  The  IoT  includes  all  things  that have a unique id and a presence in an Internet-like structure.
  • The  Internet  of  Locations  (IoL):  refers  to  all  data  that  have  a  geographical  or  geospatial  dimension.  With  the  uptake  of  mobile  devices  (e.g.  smartphones),  more  and  more events  have  location  or  movement attributes.

Note  that  the  IoC,  the  IoP,  the  IoT,  and  the  IoL  are  overlapping.  For  example,  a  place  name  on  a  webpage  or  the  location  from  which  a  tweet  was  sent.  Process  mining  aims  to  exploit  event  data in a meaningful way, for example, to provide  insights,  identify  bottlenecks,  anticipate problems, record policy violations, recommend countermeasures, and streamline  processes.  This  explains  our  focus on event data.

Process mining is an emerging discipline  providing  comprehensive  sets  of  tools  to  provide  fact-based  insights  and  to  support  process  improvements.  This  new discipline builds on process model-driven  approaches  and  data  mining.  Process  mining  provides  a  generic  collection of techniques to turn event data into  valuable  insights,  improvement  ideas,  predictions,  and  recommendations.  The  starting  point  for  any  process  mining  effort  is  a  collection  of  events commonly referred to as an event log (although events can also be stored in a  database  and  may  be  only  available  as  an event stream). A wide range of process mining  techniques  is  available  to  extract  value  and  actionable  information  from  event  data.  Process  discovery  techniques  take  an  event  log  or  event  stream  as  input  and  produce  a  process  model  without  using  any  a-priori  information.  Conformance  checking  can  be  used  to  check  if  reality,  reflected  by  the  event  data,  conforms  to  a  predefined  process  model  and  vice  versa.  Process  mining  can also be used to extend process models  with  performance-related  information, e.g., bottlenecks, waste, and costs. It is event possible to predict problems and suggest actions.

Currently,  there  are  over  25  commercial process mining tools. In fact, the adoption  of  process  mining  has  been  accelerating  in  recent  years.  Tools  like  Disco  (Fluxicon),  Celonis  Process  Mining,  ProcessGold  Enterprise  Platform,  Minit,  myInvenio,  Signavio  Process  Intelligence,  QPR  ProcessAnalyzer,  LANA  Process  Mining,  Rialto  Process,  Icris  Process  Mining  Factory,  Worksoft  Analyze  &  Process  Mining  for  SAP,  SNP  Business  Process  Analysis,  web-Methods Process Performance Manager, and  Perceptive  Process  Mining  are  now  available.  Moreover,  open  source  tools  like  ProM,  ProM  Lite,  and  RapidProM  are widely used. It is vital that event data can  be  exchanged  between  these  tools.  Several  of  these  tools  already  support  XES. For example, it is easy to exchange XES  data  between  Disco,  Celonis,  ProM,  Rialto Process, minit, and SNP.