The Best Process Mining Dissertation Award is awarded by the IEEE Task Force on Process Mining to an outstanding PhD thesis focused on the area of business process intelligence. The award is particularly dedicated to works contributing to research in the area of process mining and/or the innovative use of process mining techniques for solving practically relevant problems.

With this award, the IEEE Task Force on Process Mining wants to draw attention to excellent works by young researchers and promote the research area as a whole.

The winner of the 2014 Best Process Mining Dissertation Award is Andrea Burattin for his PhD thesis entitled Applicability of Process Mining Techniques in Business Environments.

The verdict of the Jury:

This thesis has the bold goal of democratising process mining. Specifically, it addresses the obstacles faced when trying to apply process mining techniques in business environments. Andrea identified three possible sources of problems: i) during the preparation of the data (incompatibility between required and available data) ; ii) during mining (adaptation of existing algorithms in order to take full advantage of the available information, configuration of the parameters of the mining algorithms) ; iii) during the interpretation of the mining results (complexity of the generated output). He also addressed the orthogonal issue of computation complexity for data management. In view of the identified problems, the thesis proposes solutions that can be placed in two application scenarios: the first scenario considers the classic paradigm of batch process mining, where a log is available “off-line” for mining; the second scenario introduces the novel paradigm of “online” process mining, where events are emitted by a stream with infinite capacity, thus requiring on-the-fly mining using bounded memory and time resources.

The thesis embraces the entire discovery process, from data preparation through to model inference and results presentation, offering one or more solutions to the problems identified in each of these phases, while investigating different viewpoints (e.g. control-flow vs resources, imperative vs declarative). Thus, the topic is treated extensively, both in width and depth. The specific contributions of this thesis are:

• a solution for the identification of the “case-ids” whenever this field is hidden (i.e., when it is not explicitly reported);
• a generalization of Heuristics Miner to exploit non instantaneous events;
• definition of an automatic approach for the extension of a control-flow model with social information (i.e., roles), in order to simplify the analysis of these perspectives (control-flow and resources);
• (Heuristics Miner) parameters configuration for not-expert users, exploring both the application of the Minimum Description Length principle (to balance model complexity and data explanation), and human interaction (to navigate a hierarchy of models and find the most suitable model);
• definition of two metrics for data interpretation and results evaluation: a model-to-model and a model-to-log (the latter considers models expressed in declarative language).
• two baseline approaches for online process mining, used for validation purposes;
• a general framework for defining mining algorithms (such as Online Heuristics Miner) which can be used for different kinds of streams (for example, stationary streams or streams with concept drifts);
• error bounds on Online Heuristics Miner;
• a control-flow discovery algorithm based on a well-known frequency counting algorithm (i.e., Lossy Counting) with error guarantees.

The contributions are clearly significant and innovative. Further, the thesis has high potential of generating impact both for academia and in practice. The techniques presented are both innovative and generalizable (notably, the framework for defining discovery algorithms from event streams), thus offering the research community many opportunities for further developments. Given the goal of the thesis, the impact for practice is immediate, especially for SMEs.

Despite the vast amount of techniques contributed, the technical depth of the thesis is consistently high throughout: there is a good mix of conceptualisation (e.g. architectural design), formalisation (definitions, theorems and proofs) and algorithms. All techniques have been implemented and evaluated both with synthetic and real-life data.

The Call for 2014 received two further excellent nominations.

• Claudio di Cicco: On the Mining of Artful Processess
• Jochen de Weerdt: Business Process Discovery: New Techniques and Applications

Descriptions of both nominations will be available shortly.

Below you find the call for nominations for the year 2014, which you can also download as .pdf.

The Best Process Mining Dissertation Award 2014 will be conferred by the IEEE Task Force on Process Mining to the winner during the meeting of the IEEE Task Force on Process Mining at the 12th Int. Conference on Business Process Management on 7 September in Haifa, Israel. As part of this event, the recipient will be invited to give a presentation on the main results of the thesis in a form suitable for the event and audience.

The selected thesis will also be recommended for publication as a monograph in the LNBIP series published by Springer. Further, the Dissertation Award will be accompanied by a monetary prize of 1,000 EUR.

• The call for nominations refers to PhD theses which have been successfully defended (and the PhD title awarded) in the period between 1 Jan 2012 and 31 Dec 2013. Nominations for theses that fall outside this period will not be taken into consideration.
• The thesis must be written in English.

1. A PhD thesis is nominated for the award by their supervisors. A nomination should consist of an electronic version of the thesis itself, accompanied by a brief letter (1 or 2 pages) from the supervisor describing why the thesis should be considered for the award.
2. Nominations should highlight the quality and impact of the work. The nominator may include a pointer to research work influenced by the results of the nominated thesis, and citations to the thesis or derived works, where available.
3. Submissions should be sent to the Award Organizers at the following email addresses: Dirk Fahland, Antonella Guzzo, Marcello La Rosa.

Nominations are judged by an international committee of five experts, members of the IEEE Task Force on Process Mining, based on the following criteria:

• Significance and innovation of the research contribution
• Technical depth of the research contribution, in terms of
  1. conceptualization
  2. formalization
  3. implementation
  4. evaluation
• Potential impact on academia
• Potential impact on practice
• Quality of presentation
• Quality and number of publications derived from the PhD work
• Total number of citations (incl. publications and thesis itself)

• 1 May 2014: deadline for nominations
• 1 July 2014: announcement of the result to the winner
• 7 September 2014: awards ceremony at the IEEE Task Force meeting (as part of BPM’14)

• Antonella Guzzo, University of Calabria (co-chair)
• Marcello La Rosa, QUT (co-chair)
• Boudewijn van Dongen, TU/e
• Barbara Weber, University of Innsbruck
• Diogo Ferreira, University of Lisboa

• Dirk Fahland, Eindhoven University of Technology, The Netherlands https://www.win.tue.nl/~dfahland/
• Antonella Guzzo, Università della Calabria, Italy http://si.deis.unical.it/~guzzo/
• Marcello La Rosa, Queensland University of Technology, Australia http://marcellolarosa.com