First International Workshop on Data Management and Analytics for Semi-Structured Business Processes (dma4sp 2011)
In conjunction with International Conference on Data Engineering (ICDE) 2011
Hannover, Germany
April 11, 2011

Abstract and Goals
Human-centric, ad-hoc, and dynamic semi-structured processes are common in today?s business enterprises. The lifecycle of such processes is not necessarily driven by a formal process model and the execution of such processes may be very dynamic. Such processes may execute on heterogeneous platforms and involve the interchange of diverse documents and artifacts (E.g. emails, images, PDF attachments, videos, audio files, chat transcripts). Key problems that arise as a result of the proliferation of such processes include how to integrate unstructured data (E.g. emails, execution logs, conversation transcripts, social media) for semi-structured business processes from disparate sources and heterogeneous platforms (E.g. social networking platforms, heterogeneous databases, distributed legacy workflow engines). Related challenges include mining data of such business processes to enable modeling, optimization, prediction, collaboration and community management, network analysis and visualization. The goal of this workshop is to investigate novel algorithms, systems, platforms to address such problems, discuss application case-studies that showcase the need for such solutions, and study emerging applications where such solutions may already have been deployed.

Submission deadline: November 1, 2010
Notification of acceptance: December 1, 2010
Camera ready papers due: January 1, 2011
Workshop: April 11, 2011

Complete call-for-papers, submission instructions and program committee details on workshop website: https://sites.google.com/site/dma4sp/

Workshop Co-Chairs
Charu Aggarwal , IBM T. J. Watson Research Center, charu@us.ibm.com
Geetika T. Lakshmanan , IBM T. J. Watson Research Center, gtlakshm@us.ibm.com
Elke Rundensteiner , WPI, rundenst@sequoia.wpi.edu

Questions? Contact gtlakshm@us.ibm.com

Detailed workshop overview

Human-centric, ad-hoc, and dynamic semi-structured processes are common in today?s business enterprises. The lifecycle of such processes may not be driven by a formal process model and the execution of such processes may be very dynamic. Such processes may execute on heterogeneous platforms and involve the interchange of diverse documents and artifacts (emails, images, PDF attachments, videos, audio files, chat transcripts). Such processes may also evolve with the introduction of new social communication platforms such as Chatter, Facebook and Twitter and different social media (blogs, wikis, social bookmarks, video sharing, presentation sharing, etc).

Consider a case in the insurance industry. An insurance case is managed by a chain of human employees in an insurance company. A given case may not fall in the realm of any single formal representation of the insurance process. The case may be handled over the phone with conversations between the customer and the case manager and various paper documents or emails may be exchanged that are related to the case. Employees may also discuss the case over internal (company firewall protected) social networking platforms. The case may be stored in a database, and some case history could be maintained. Using some basic case history, how can individual case instances be tracked, integrated and modeled as a formal business process? Furthermore, how can the community and social networks involved in such processes be mined and represented?

Collaborative processes in healthcare are another example of semi-structured business processes. The admission of a patient to a hospital could be viewed as an instance of a semi-structured process. The basic workflow of a hospital is to handle these cases. Activities in such a workflow could include all kinds of treatments, operations. Electronic forms have to be filled out, emails are exchanged, and different medical professionals including doctors, nurses, staff and specialists are involved. In such processes while medical staff can follow formal guidelines for procedures, they may not do so for unique situations. How can such processes be tracked, analyzed, formally modeled, and how can predictive models for such processes be constructed to model future outcomes?

In the banking industry while formal structured processes model business processes such as mortgage applications, employees can make decisions and take actions that fall outside the realm of a formal process. How can such actions be tracked and modeled? How can financial fraud be detected through analytics on data for financial case history?

Key problems that arises in this space include how to integrate unstructured data (E.g. emails, execution logs, conversation transcripts, social media) for semi-structured business processes from disparate sources and heterogeneous platforms (E.g. social networking platforms, heterogeneous databases, distributed legacy workflow engines). Related challenges include mining data of such processes to enable modeling, optimization, prediction, collaboration and community management, and network analysis. The goal of this workshop is to investigate novel solutions to these problems.

The data mining and business process management communities are currently quite separate. For example, the community in the field of semi-structured and unstructured data works on mining documents of different kinds such as free form xml or text data. Similarly, the image and video processing communities work with data of various kinds. The business process management community is a different community of researchers, which tends to work independently. Researchers in each of these communities think of similar problems related to data and process management. Furthermore both communities experience equal impact from recent rapid advances in the way data evolves and is exchanged brought on by the proliferation of social network and communication platforms and different social media. This workshop intends to bring researchers together from both communities to engage in an exchange of ideas to further collaborative research in the two fields on problems of common interest. Such a fusion is likely to lead to a learning experience for both communities.

Papers on, but not limited to, the following topics under the scope of this workshop are encouraged:
- Predictive modeling
- Process graph or workflow graph mining and optimization
- Analysis of social media and collaborative tools integrated into business management platforms
- Learning algorithms
- Data integration, event correlation and event processing
- Semantic interpretation and analysis of heterogeneous and semi-structured data
- Meta data management and interoperability
- Community or social network detection and management
- Monitoring data from business management platforms and social media platforms
- Visual interfaces for mined process data and visual interactions with mined process data