Recognizing and Splitting Conditional Sentences for Automation of Business Processes Management

Proceedings of Recent Advances in Natural Language Processing, pages 1490–1497 Sep 1–3, 2021. https://doi.org/10.26615/978-954-452-072-4_167 1490 Recognizing and Splitting Conditional Sentences for Automation of Business Processes Management Ngoc Phuoc An Vo, Irene Manotas, Octavian Popescu, Algimantas ˇCerniauskas*, Vadim Sheinin IBM Research, Yorktown Heights, US *IBM Client Innovation Center Baltic, Vilnius, Lithuania {ngoc.phuoc.an.vo, irene.manotas, algimantas.cerniauskas}@ibm.com {o.popescu, vadims}@us.ibm.com Abstract Business Process Management (BPM) is the discipline which is responsible for manage- ment of discovering, analyzing, redesign- ing, monitoring, and controlling business pro- cesses. One of the most crucial tasks of BPM is discovering and modelling business processes from text documents. In this paper, we present our system that resolves an end-to- end problem consisting of 1) recognizing con- ditional sentences from technical documents, 2) finding boundaries to extract conditional and resultant clauses from each conditional sentence, and 3) categorizing resultant clause as Action or Consequence which later helps to generate new steps in our business process model automatically. We created a new dataset and three models to solve this problem. Our best model achieved very promising results of 83.82, 87.84, and 85.75 for Precision, Recall, and F1, respectively, for extracting Condition, Action, and Consequence clauses using Exact Match metric. 1 Introduction Business processes which serve as basis for many (if not all) companies are usually stored as unstruc- tured data, especially as text documents (Blumberg and Atre, 2003). They can reflect all organiza- tion tasks and activities in order to provide services. Therefore, these documents could be converted into process models which allows to discover, analyze, redesign, monitor and control these business pro- cesses (M. Dumas, 2018). The discipline which is responsible for management of this life cycle is known as Business Process Management (BPM). Today, this well known practice is very inter- disciplinary and combines: organizational theory, management, and even computer science. One of the sub-areas of computer science is Natural Lan- guage Processing (NLP) and due to large number of freely available NLP tools such as taggers, parsers, and the lexical database WordNet, NLP is applied in a multitude of domains and BMP is no excep- tion. The application of NLP techniques could be used to extract useful information from textual doc- uments to infer a process model, and also to extract information from the process model to facilitate visual process analysis (Leopold, 2013). Process modelling is a very complex and time- consuming task. In this work we mainly focused on the analysis of extracted information from text documents in order to have more accurate pro- cess models. Given a list of extracted sentences from technical documents, the contributions of our work consists of the following: 1) classifying if a sentence is a conditional sentence, 2) identifying and extracting the conditional and resultant clauses from a conditional sentence, and 3) categorizing the extracted resultants. Section 2 introduces busi- ness context and application for our tasks. Section 3 explains how we defined the tasks and created the new dataset TechCond towards solving the tasks. Section 4 presents our methods and models for …

Content truncated to 500 words