R. Agrawal, D. Gunopulos, F. Leymann: "Mining process models from workflow logs". In: Proc. of the Intl. Conf. on Extending Database Technology EDBT'98, Valencia, Spain, March 3-8, 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....modeling language for workflow processes and the definition of a workflow log. Then we present a new technique for process mining. Finally, we conclude the paper by summarizing the main results and pointing out future work. 2. Related Work and Preliminaries The idea of process mining is not new [3, 4, 5, 8]. However, most results are limited to sequential behavior. Cook and Wolf extend their work to concurrent processes in [5] They also propose specific metrics (entropy, event type counts, periodicity, and causality) and use these metrics to discover models out of event streams. This approach is ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining process models from workflow logs. In the proceedings of the Sixth International Conference on Extending Database Technology, pages 469-483, 1998.

....4) and the implementation of this technique in the workflow mining tool Little Thumb (Section 5) In Section 6 we present our experimental results. Finally, we conclude the paper by summarizing the main results and pointing out future work. 2. Related work The idea of process mining is not new [6,8,9,10,14,15,16,17,21,22,24,25,26]. Cook and Wolf have investigated similar issues in the context of software engineering processes. In [8] they describe three methods for process discovery: one using neural networks, one using a purely algorithmic approach, and one Markovian approach. The authors consider the latter two the most ....

....a set of dependency relations between events. In [10] Cook and Wolf provide a measure to quantify discrepancies between a process model and the actual behavior as registered using event based data. The idea of applying process mining in the context of workflow management was first introduced in [6]. This work is based on workflow graphs, which are inspired by workflow products such as IBM MQSeries workflow (formerly known as Flowmark) and InConcert. In this paper, two problems are defined. The first problem is to find a workflow graph generating events appearing in a given workflow log. The ....

[Article contains additional citation context not shown here]

R. Agrawal, D. Gunopulos, and F. Leymann. Mining process models from workflow logs. In the proceedings of the Sixth International Conference on Extending Database Technology, pages 469-483, 1998.

....mining. For this purpose, the log is used for discovering the business rules and the workflow models which lead to the execution patterns which are observed. Data mining in the context of workflow logs to discover information about the workflow instances of various kinds is addressed e.g. in [3, 5, 9, 19, 20]. Di#erent methods have emerged, targeting di#erent kinds of data, such as workflow control structure, resource allocation, time consumption parameters and so on. The 7. CONCLUSIONS Fig. 7. example of analysis 2 detail specialized methods take advantage of the nature of the specific data to ....

Rakesh Agrawal, Dimitrios Gunopulos, and Frank Leymann. Mining process models from workflow logs. In Hans-Jorg Schek, Felix Saltor, Isidro Ramos, and Gustovo Alonso, editors, Advances in Database Technology - EDBT'98, 6th International Conference on Extending Database Technology, Valencia, Spain, March 23-27, 1998.

....probabilities. 5. Related work Although there is a common agreement that logging and analysis of workflow executions are important tasks in workflow management [10, 27, 24, 25, 21, 20, 9] little work has been done in the area of analyzing and mining the histories of workflows. Agrawal et al. [1, 2] consider the problem to generate a workflow model from a log of executions produced by a preexisting system which uses a different (usually less formal) representation. the developed models differ due to the different application contexts. Whereas Agrawal et al. aim at a model which forms the ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining process models from workflow logs. In Proc. 6th Intl. Conference on Extending Database Technology (EDBT'98), Valencia, Spain, March 1998.

....I K J L Fig. 1. A process model for the log shown in Table 1 A parallel execution of tasks H and G means that they can appear in any order. The idea of discovering models from process logs was previously investigated in contexts such as software engineering processes and workflow management [2 9]. Cook and Wolf propose three methods for process discovery in case of software engineer processes: a finite state machine method, a neural network and a Markov approach [3] Their methods focus on sequential processes. Also, they have provided some specific metrics for detection of concurrent ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process models from Workflow Logs. In Sixth International Conference on Extended Database Technology, pg. 469-483, 1998.

....is an access method s ability to sustain high insert update rates. This requirement arises, for example, in very large data warehouses [CACM98, Wor98] in scientific databases that are fed by automatic instruments [MSS95] or in workflow management systems for keeping workflow histories [AGL98]. Also, many banking and stock market applications exhibit such characteristics. For example, consider the management of stock portfolios in a large bank. For each portfolio, all buy and sell orders must be tracked. Based on this data, in addition to querying the current contents of a portfolio, ....

R. Agrawal, D. Gunopulos, F. Leymann, Mining Process Models from Workflow Logs, Proc. Int. Conf. on Extending Database Technology (EDBT), Valencia, 1998 large95] L. Arge, The Buffer Tree: A New Technique for Optimal I/O Algorithms, Proc. Int. Workshop on Algorithms and Data Structures, Springer, 1995

....following. Acquiring workflow models and adapting them to changing requirements is a time consuming and error prone task, because process knowledge is usually distributed among many different people and because workflow modeling is a difficult task, that needs to be done by modeling experts (see [1], 5] or [9] Thus there has been interest in applying machine learning techniques to induce workflow models from traces of manually enacted workflow instances. The learning algorithms, we are aware of, share some restrictions, that may prevent them from being used in practice. They either apply ....

....instances. The learning algorithms, we are aware of, share some restrictions, that may prevent them from being used in practice. They either apply grammatical inference techniques and are restricted to sequential workflows [5] 9] or they allow concurrency but require unique activity nodes [1] [6] 2 Definitions In the following we define the terms workflow model and workflow instance. This is essential for a description of the induction task. A workflow model is a formal explicit representation of a business process, describing how this process is (or should be) performed. It ....

[Article contains additional citation context not shown here]

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. In Proc. of the sixth International Conference on Extending Database Technology (EDBT), 1998.

....the description such obtained really the optimal one Therefore, it is important that the analy sis results in a canonical form. With such a canonical deterministic finite automaton (dfa) one aims for the following properties, which taken together determine the quality of the resulting automaton [1]: Completeness: The dfa should preserve all transitions which can be found in the trace data. All sequences in the log must be generated by the dfa. Irredundancy: No transition of the dfa should be spurious. This property should prohibit incorrect transitions in the resulting automaton. ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. In H.-J. Schek, F. Saltor, I. Ramos, and G. Alonso, editors, 6 th International Conference on Extending Database Technology - EDBT'98, Valencia, Spain, volume 1377 of Lecture Notes in Computer Science, pages 469 -- 483. Springer, March 1998.

.... kinds of analysis [BCDS01] and the application of data mining techniques for understanding, predicting, and preventing exceptions [GCDS01] Another interesting area of research is business process discovery, where the goal is to learn the structure of a business process from workflow log data [AGL98]. This would be especially useful for semi structured and unstructured business activities where the process is not defined a priori. Such a process is usually implemented via rules triggered by events such as the start and completion of tasks. In some situations, there may be a latent structure ....

....that a complete business process that spans activities in different enterprises will be easily defined and agreed to by all the participants. In such situations, it may be possible to learn the underlying process by analyzing the sequences of interactions among the participants. Although [AGL98] made a promising start in this direction, process discovery is very much an open research issue. 6. Summary Business process integration and automation have become high priorities for enterprises to achieve operational efficiency. With the burgeoning of e commerce, there is a renewed interest ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. Proc. of the Sixth International Conference on Extending Database Technology (EDBT), Valencia, Spain, 1998.

....equally convincing. The algorithm has some deficiencies, when long concurrent threads of activities are present. In this case it is very likely to detect incorrect dependencies. We are thinking about using additional background knowledge for the detection of dependencies. 7 Related Work In [ Agrawal et al. 1998 ] an approach for inducing workflow models from workflow instances is presented. This approach is based on the induction of directed graphs. It is restricted to the problem class 3 and very similar to our approach for this class. Another approach called RAP is presented in [ Bocionek and Mitchell, ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. In Proc. of the sixth International Conference on Extending Database Technology (EDBT), 1998.

....if T 2 read data from the blackboard that was written by T 1 . This dependency information can easily retrieved, since usually a workflow system logs all information pertaining to process execution in its execution log for legal reasons and to facilitate process optimization based on statistics [1]. We denote with DEP (t) the transitive closure of DEPALL (t) As a shortcut, we define DEP (T; t) as the restriction of DEP (t) on those tasks causally dependent on T , i.e. DEP (T; t) contains all tasks transitively causally dependent on T . The goal of the partial rollback ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining process models from workflow logs. In EDBT 98, 1998.

....effectiveness in understanding it and in making the correct changes for sound improvement of the system. While our work in discovering concurrency from event traces appears to be novel, there certainly has been related work in understanding distributed, concurrent systems. ffl Agrawal et al. [1] investigate producing activity dependency graphs from event based workflow logs. The logs already identify the partial ordering of concurrent, time spanning activities, and they are concerned with producing correct and minimal graphs. There is no notion of identifying synchronization points ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. Technical Report (draft technical report), IBM, September 1997.

....if T 2 read data from the blackboard that was written by T 1 . This dependency information can easily retrieved, since usually a workflow system logs all information pertaining to process execution in its execution log for legal reasons and to facilitate process optimization based on statistics [AGL98] We denote with DEP (t) the transitive closure of DEPALL (t) As a shortcut, we define DEP (T ; t) as the restriction of DEP (t) on those tasks causally dependent on T , i.e. DEP (T ; t) contains all tasks transitively causally dependent on T . The goal of the partial rollback ....

R. Agrawal, D. Gunopulos, and F. Leymann. Mining process models from workflow logs. In EDBT 98, 1998.

No context found.

R. Agrawal, D. Gunopulos, F. Leymann: "Mining process models from workflow logs". In: Proc. of the Intl. Conf. on Extending Database Technology EDBT'98, Valencia, Spain, March 3-8, 1998.

No context found.

Agrawal, R., Gunopulos, D. & Leymann, F. (1998), Mining Process Models from Workflow Logs, in on Extending Database Technology (EDBT'98)', Valencia, Spain.

No context found.

R. Agrawal, D. Gunopulos, and F. Leymann. Mining process models from workflow logs. In Proc. 6th Int. Conf. on Extending Database Technology (EDBT'98), pages 469--483, 1998.

No context found.

R. Agrawal, D. Gunopulos, and F. Leymann. Mining process models from workflow logs. In Proc. 6th Int. Conf. on EDBT'98, 469--483, 1998.

No context found.

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. In Sixth International Conference on Extending Database Technology, pages 469--483, 1998.

No context found.

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. In Proceedings of the sixth International Conference on Extending Database Technology (EDBT), 1998.

No context found.

Rakesh Agrawal, Dimitrios Gunopulos, Frank Leymann. 1998. "Mining Process Models from Workflow Logs," Proceedings of the Sixth International Conference on Extending Database Technology (EDBT).

No context found.

R. Agrawal, D. Gunopulos, and F. Leymann. Mining Process Models from Workflow Logs. In Sixth International Conference on Extending Database Technology, pages 469--483, 1998.

No context found.

R. Agrawal, D. Gunopulos, F. Leymann, Mining process models from workflow logs, Lecture Notes in Computer Science 1377 (1998) 469--483.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC