Extracting Annotations from Textual Descriptions of Processes

Abstract

Organizations often have textual descriptions as a way to document their main processes. These descriptions are primarily used by the company's personnel to understand the processes, specially for those ones that cannot interpret formal descriptions like BPMN or Petri nets. In this paper we present a technique based on Natural Language Processing and a query language for tree-based patterns, that extracts annotations describing key process elements like actions, events, agents/patients, roles and control-flow relations. Annotated textual descriptions of processes are a good compromise between understandability (since at the end, it is just text), and behavior. Moreover, as it has been recently acknowledged, obtaining annotated textual descriptions of processes opens the door to unprecedented applications, like formal reasoning or simulation on the underlying described process. Applying our technique on several publicly available texts shows promising results in terms of precision and recall with respect to the state-of-the art approach for a similar task.

Improving the Extraction of Process Annotations from Text with Inter-sentence Analysis

“…In this paper we describe an extension to the approach presented in [9]. This subsection summarizes the basic original approach, and following subsections provide details on the added extensions, which mainly consist of the extraction of relations between actions or conditions in different sentences, as well as an extended evaluation covering not only entities and actions, but also relations.…”

“…In [9] we presented a proposal to extract Business Process elements (entities, actions, conditions, events, and relations) from a process textual description.…”

“…Patterns used in [9] for relation extraction summarized in Sect. 4.1 aimed to capture relations between two activities/events mentioned in the same sentence.…”

“…Conflicts between activities in the same sentence are detected using patterns described in [9]. The following patterns deal with conflicts between activities in different sentences.…”

“…In this paper we significantly expand the techniques and results recently presented in [9], where we described robust tree-based patterns to be queried over the dependency trees arising from the NLP analysis of the textual descriptions. Patterns in [9] where only applicable in the context of a single sentence, which made our approach unable to extract inter-sentence relationships. The contribution of this paper is therefore the extension of our previous contribution with a more general set of patterns, resulting in a significant boost in the recall of the original framework (see experiment results in Sect.…”