Abstract | ||
---|---|---|
The increasing number of process models in an organization has led to the development of process model repositories, which allow to efficiently and effectively manage these large number of models. Searching process models is an inherent feature of such process repositories. However, the effectiveness of searching depends upon the accuracy of the underlying matching technique that is used to compute the degree of similarity between query-source process model pairs. Most of the existing matching techniques rely on the use of labels, structure or execution behavior of process models. The effectiveness of these techniques is, however, quiet low and far from being usable in practice. In this paper, we address this problem and propose the use of a combination of textual descriptions of process models and text matching techniques for process matching. The proposed approach is evaluated using the established metrics, precision, recall and F-1 score. The results show that the use of textual descriptions is slightly more effective than activity labels. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1007/978-3-319-48393-1_14 | Lecture Notes in Business Information Processing |
Field | DocType | Volume |
QUIET,USable,F1 score,Degree of similarity,Systems engineering,Pattern recognition,Computer science,Process modeling,Artificial intelligence,Recall,Machine learning | Conference | 267.0 |
ISSN | Citations | PageRank |
1865-1348 | 1 | 0.35 |
References | Authors | |
14 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Maria Rana | 1 | 1 | 0.35 |
Khurram Shahzad | 2 | 165 | 25.77 |
Muhammad Adeel | 3 | 27 | 11.84 |
Henrik Leopold | 4 | 504 | 36.19 |
Umair Babar | 5 | 1 | 0.35 |