Ontology-based Business Process Chaining in Heterogeneous Systems

Abstract

Large companies deal with complex business processes that span across multiple applications. Process Mining (PM) across the applications is not straight forward due to di ering content, format, and context of the event logs. Ontology-based techniques can capture important metadata about applications and processes. The proposed work aims to use the relationship between metadata to identify processes that run across applications. We intend to apply the domain context for extending similarity measures within and across ontologies.

Keywords

• Business Knowledge Representation
• Domain-Speci c On- tology
• Event Logs
• Knowledge Graph
• Natural Language Processing
• Ontology Similarity Measures
• Process Mining

-

2. Modeling process drifts and overlaps of dynamically evolving processes. 3. Mining from diverse logs that lack structural or transitional relationships 4. Automating ontology enrichment, validation and downstream rule updates. 5. Using domain context, Key Performance Indicators and crossontology relations for PM and ontology pruning 6. Extending distance metrics of ontologies to include domain and implementation-speci c factors 2

Approach

Ontology-based techniques use metadata in the form of domain concepts or process descriptions [ 9 ][ 10 ] when identifying and sequencing tasks. Our research is primarily a software application problem that uses semantic web and ontology techniques to help process mining for complex processes. We propose to merge event-log based PM techniques with ontology creation and distance metricsbased concept association techniques to discover processes across applications. We further propose expanding the syntactic, structural and semantic similarity measurement techniques [ 11 ] to establish relationships between concepts within and across ontologies.

Figure 1 captures the ow of our research. Our research involves the analysis of event logs from each application separately using available PM techniques. We intend to create granular ontologies for each application combining pieces of discovered processes, domain concepts, application-speci c concepts, and heuristics about the process. We plan to apply similarity measures to identify relationships that span across the ontologies and derive process associations across heterogeneous logs. 3

Methodology

The proposed research will adapt Design Science Research (DSR) [ 12 ] for applying ontology-based techniques to mine complex processes. Per the ED Process [ 13 ], we propose to follow an iterative approach consisting of design, prototype, and evaluation for making our research available in a toolkit. We will use the mixed-methods research methodology for data collection from event logs (quantitative) and heuristic information from domain experts (qualitative). We will further use PM speci c data preprocessing techniques [ 3 ] to cleanse, remove noise, and x errors in data such as incomplete traces, missing events, mashed processes and many more. 4

Intended solution

Figure 2 presents our solution as a combination of PM and Ontology techniques. The research project will discover processes from heterogeneous event logs in the chosen domain. We plan to use ProM as the PM tool due to its wide selection of discovery algorithms [ 14 ]. We will create ontologies to represent the domain context of the applications. We have shortlisted Protege as the tool for ontology creation and analysis, due to its domain-friendly support for creating knowledge models [ 15 ].

Ontology-based technique have proven useful in PM [ 16 ][ 17 ]. Similarity measures is a common technique to nd the relationship between concepts across ontologies [ 18 ]. We propose to devise PM speci c distance and similarity measures to establish the relationship between the concepts within and across ontologies. We further plan to use a combination of concept-level and ontology-level relationships to correlate processes in heterogeneous applications. 5

Contributions to BPM research

The issue of PM across heterogeneous applications is not fully explored, which con rms our identi ed research gap. Our research focuses on arriving at similarity measures that work across multiple domain-speci c ontologies. We propose to use those measures to discover and analyse complex processes that ow across heterogeneous applications. The techniques developed will also aid organisations [ 19 ] in change management, predicting disruptions and achieving better optimisation across their business. We plan to create a toolkit and sample implementations that would help future researchers in creating multiple domain-speci c ontologies. 6

Project status and challenges

Current state of the research project is presented in Table 1. Phase Task Status Literature survey Survey across PM, ontology techniques and intel- Completed and Analysis ligent decision systems.

Data Preparation Identify, analyse and preprocess datasets. In Progress Process Mining Baseline using standard PM techniques and tools Pending Ontology Creation Creating and re ning domain-speci c ontologies Pending with process-speci c concepts Establishing Simi- Exploring existing similarity measures and extend- Pending larity Measures ing them to use PM and domain-speci c features Evaluation Evaluating usefulness of ontologies and similarity Pending measures using available techniques

Develop- Design and Implement Toolkit Pending Our work of applying ontology and distance measures to discover processes can help detect process drifts. It can also aid organisations in identifying overlaps and hidden relationships of processes across various applications. By adding KPI information, techniques developed in our research can be useful in automating downstream rule updates. 6.1

Glossary of Terms 1. Business Process: A collection of tasks or activities performed by people or equipment to achieve a concrete goal in an organisation. 2. Business Process Analysis: Methodology to understand the health of di erent operations in a business. 3. Business Process Chaining: Establishing the relationship between parts of processes across system and application boundaries. 4. Domain-Speci c Ontology: An ontology or series of ontologies that represent a set of concepts and relationships speci c to the domain. 5. Event Logs: Digitized data about events executed by IT applications. 6. Heterogeneous Applications: Information Technology (IT) applications implemented using di erent technologies, following di erent standards and formats for data and logic. 7. Intelligent Decision System (DSS): A class of computerised information system that supports business and organisational decision-making activities through rules and learning algorithms. 8. Intelligent Process Automation: An emerging set of new technologies that combines fundamental process redesign with robotic process automation and machine learning. It is a suite of business-process improvements and next-generation tools that assists the knowledge worker by removing repetitive, replicable, and routine tasks. 9. Ontology: Ontology is the formal representative of concepts and their relationships. This enables computers and humans to interpret semantic relationships among the concepts and infers implicit knowledge [ 20 ]. 10. Process Mining (PM): Approach and techniques to discover, monitor and improve real processes (i.e., not assumed processes) by extracting knowledge from event logs readily available in today's (information) systems. PM includes (automated) process discovery (i.e., extracting process models from an event log), conformance checking (i.e., monitoring deviations by comparing model and log), social network/organisational mining, automated construction of simulation models, model extension, model repair, case prediction, and history-based recommendations [ 19 ].

References

1. Adisa Delic , Sabina Donlagic Alibegovic, and Mersiha Mesanovic . The role of the process organizational structure in the development of intrapreneurship in large companies . Nase gospodarstvo/Our economy, 62 ( 4 ): 42 { 51 , 2016 .

2. Rodney McAdam and Daniel McCormack . Integrating business processes for global alignment and supply chain management . Business Process Management Journal , 2001 .

3. RP Jagadeesh Chandra Bose , Ronny S Mans, and Wil MP van der Aalst . Wanna improve process mining results? In 2013 IEEE symposium on computational intelligence and data mining (CIDM) , pages 127 { 134 . IEEE, 2013 .

4. XES Working Group et al. IEEE standard for eXtensible Event Stream (XES) for achieving interoperability in event logs and event streams . IEEE Std 1849 , pages 1 { 50 , 2016 .

5. Suriadi Suriadi , Robert Andrews, Arthur HM ter Hofstede, and Moe Thandar Wynn . Event log imperfection patterns for process mining: Towards a systematic approach to cleaning event logs . Information Systems , 64 : 132 { 150 , 2017 .

6. Marcos Rivas Pen~a and Sussy Bayona-Ore. Process mining and automatic process discovery . In 2018 7th International Conference On Software Process Improvement (CIMPS) , pages 41 { 46 . IEEE, 2018 .

7. Edgar Tello-Leal , Jorge Roa, Mariano Rubiolo, and Ulises M Ramirez-Alcocer . Predicting activities in business processes with lstm recurrent neural networks . In 2018 ITU Kaleidoscope: Machine Learning for a 5G Future (ITU K), pages 1{7 . IEEE, 2018 .

8. Mahmoud AbdEllatif , Marwa Salah Farhan, and Naglaa Saeed Shehata. Overcoming business process reengineering obstacles using ontology-based knowledge map methodology . Future Computing and Informatics Journal , 3 ( 1 ):7{ 28 , 2018 .

9. Ji-In Nam , Pawan Nagwani, Sae-Bom Jang , Young-Bin Shin , and Ho Jin . Ontology-based intelligent home assistance system . In 2016 IEEE International Conference on Consumer Electronics (ICCE) , pages 121 { 122 . IEEE, 2016 .

10. TITI Sondes, Hadda BEN ELHADJ, and Lamia CHAARI . An ontology-based healthcare monitoring system in the internet of things . In 2019 15th International Wireless Communications & Mobile Computing Conference (IWCMC) , pages 319 { 324 . IEEE, 2019 .

11. Oussama El Hajjamy, Larbi Alaoui, and Mohamed Bahaj . Semantic integration of heterogeneous classical data sources in ontological data warehouse . In Proceedings of the International Conference on Learning and Optimization Algorithms: Theory and Applications , pages 1 {8 , 2018 .

12. Roel J Wieringa. Design science methodology for information systems and software engineering. Springer, 2014 .

13. Andrew Kusiak . Engineering design: products, processes, and systems . Academic Press, Inc., 1999 .

14. HMW Verbeek, Wil MP van der Aalst , and Jorge Munoz-Gama. Divide and conquer: A tool framework for supporting decomposed discovery in process mining . The Computer Journal , 60 ( 11 ): 1649 { 1674 , 2017 .

15. Usha Yadav , Gagandeep Singh Narula, Neelam Duhan, Vishal Jain, and BK Murthy . Development and visualization of domain speci c ontology using protege . Indian Journal of Science and Technology , 9 ( 16 ): 1 { 7 , 2016 .

16. AK Alves De Medeiros , Carlos Pedrinaci, Wil MP Van der Aalst , John Domingue, Minseok Song, Anne Rozinat, Barry Norton, and Liliana Cabral . An outlook on semantic business process mining and monitoring . In OTM Confederated International Conferences" On the Move to Meaningful Internet Systems" , pages 1244 { 1255 . Springer, 2007 .

17. Diego Calvanese, Marco Montali, Alifah Syamsiyah, and Wil MP van der Aalst . Ontology-driven extraction of event logs from relational databases . In International Conference on Business Process Management , pages 140 { 153 . Springer, 2016 .

18. Wenlong Lu , Yuchu Qin, Qunfen Qi, Wenhan Zeng, Yanru Zhong, Xiaojun Liu, and Xiangqian Jiang . Selecting a semantic similarity measure for concepts in two di erent cad model data ontologies . Advanced Engineering Informatics , 30 ( 3 ): 449 { 466 , 2016 .

19. Wil Van Der Aalst . Process mining: Overview and opportunities . ACM Transactions on Management Information Systems (TMIS) , 3 ( 2 ):1{ 17 , 2012 .

20. Ciza Thomas. Ontology in Information Science. BoD{Books on Demand , 2018 .

11. eXtensible Event Stream (XES): An IEEE speci cation for a tag-based language to capture event logs and event streams . Approved in Nov 2016 [ 4], the speci cation achieves interoperability in event logs to enable easier process discovery and analysis .