Academia.eduAcademia.edu

Outline

Title
Abstract
Figures
Related Work
Background
Flow Analysis
Overview
Obtaining the Flow Analysis Formulas
Datasets
Experimental Setup
Data Split
Evaluation Results
Conclusion and Future Work
References
All Topics
Computer Science

Predicting process performance: A white-box approach

Profile image of Marlon DumasMarlon Dumas

2017

Abstract

Predictive business process monitoring methods exploit historical process execution logs to provide predictions about running instances of a process. These predictions enable process workers and managers to preempt performance issues or compliance violations. A number of approaches have been proposed to predict quantitative process performance indicators for running instances of a process, including remaining cycle time, cost, or probability of deadline violation. However, these approaches adopt a black-box approach, insofar as they predict a single scalar value without decomposing this prediction into more elementary components. In this paper, we propose a white-box approach to predict performance indicators of running process instances. The key idea is to first predict the performance indicator at the level of activities, and then to aggregate these predictions at the level of a process instance by means of flow analysis techniques. The paper develops this idea in the context of predicting the remaining cycle time of ongoing process instances. The proposed approach has been evaluated on real-life event logs and compared against several baselines.

Figures (19)
FIGURE 1 Core BPMN elements 27. synchronization of the multiple incoming paths 27. A process model can be decomposed into process fragments. A pro-
FIGURE 1 Core BPMN elements 27. synchronization of the multiple incoming paths 27. A process model can be decomposed into process fragments. A pro-
TABLE 1 Extract of an event log. A wide range of process automated process discovery algorithms have been proposed in the literature 2’, with Petri nets and BPMN being the most common output model formats. processes, the waiting time makes up a considerable proportion of the
TABLE 1 Extract of an event log. A wide range of process automated process discovery algorithms have been proposed in the literature 2’, with Petri nets and BPMN being the most common output model formats. processes, the waiting time makes up a considerable proportion of the
FIGURE 2 Typical process model patterns: sequential (a), XOR-block (b), AND-block (c) and rework loop (d). that the rework fragment will be executed follows a geometric distribu- tion with the expected value 1/(1 — r). Thus, the average cycle time of the fragment in this case is:
FIGURE 2 Typical process model patterns: sequential (a), XOR-block (b), AND-block (c) and rework loop (d). that the rework fragment will be executed follows a geometric distribu- tion with the expected value 1/(1 — r). Thus, the average cycle time of the fragment in this case is:
FIGURE 3 Overview of the proposed approach. process, generally we need to discover the model from event logs. For
FIGURE 3 Overview of the proposed approach. process, generally we need to discover the model from event logs. For
FIGURE 4 Backtracking algorithm (taken from*?). backtracks to the parent node to traverse other branches.
FIGURE 4 Backtracking algorithm (taken from*?). backtracks to the parent node to traverse other branches.
FIGURE 5 Example process model. Highlighted is the current marking
FIGURE 5 Example process model. Highlighted is the current marking
FIGURE 6 Unfolding the rework loop of F
FIGURE 6 Unfolding the rework loop of F
TABLE 3 Training set to predict the branching probabilities of X32 (single model). 1.6.2 | Feature encoding with multiple predictive models In this work, we will use the former approach, as we are mostly inter ested in predictions in the early stages of a process evolution wher many process instances have not finished yet, so there is still sufficien amount of data to train the predictive models. In this work, we will use the former approach, as we are mostly inter-
TABLE 3 Training set to predict the branching probabilities of X32 (single model). 1.6.2 | Feature encoding with multiple predictive models In this work, we will use the former approach, as we are mostly inter ested in predictions in the early stages of a process evolution wher many process instances have not finished yet, so there is still sufficien amount of data to train the predictive models. In this work, we will use the former approach, as we are mostly inter-
TABLE 2 Training set to predict the cycle time of F' (single model).
TABLE 2 Training set to predict the cycle time of F' (single model).
TABLE 6 Statistics of the datasets used in the experiments.
TABLE 6 Statistics of the datasets used in the experiments.
TABLE 8 Weighted average MAE over all prefixes.
TABLE 8 Weighted average MAE over all prefixes.
TABLE 7 Learning parameters of XGBoost that have been tuned.
TABLE 7 Learning parameters of XGBoost that have been tuned.
FIGURE 8 Average ranking of the evaluated methods over all datasets. Error bars indicate the 95% confidence interval. FIGURE 7 Prediction accuracy (measured in terms of MAE) across different prefix lengths.
FIGURE 8 Average ranking of the evaluated methods over all datasets. Error bars indicate the 95% confidence interval. FIGURE 7 Prediction accuracy (measured in terms of MAE) across different prefix lengths.
TABLE 9 MAE of cycle time predictions of individual activities and their actual mean cycle times (in days). TABLE 10 Predicted and average values of cycle times and branching probabilities for hd? (SWKD).
TABLE 9 MAE of cycle time predictions of individual activities and their actual mean cycle times (in days). TABLE 10 Predicted and average values of cycle times and branching probabilities for hd? (SWKD).
Execution Times
Execution Times
FIGURE 10 Normalized standard deviation of MAE over all prefixes FIGURE 9 Normalized MAE averaged over all prefixes and all datasets.
FIGURE 10 Normalized standard deviation of MAE over all prefixes FIGURE 9 Normalized MAE averaged over all prefixes and all datasets.
FIGURE 11 A process model of the Hospital log. Current marking of hd? (SWKD) and its predicted future path are highlighted.
FIGURE 11 A process model of the Hospital log. Current marking of hd? (SWKD) and its predicted future path are highlighted.

Related papers

Comparing and Combining Predictive Business Process Monitoring Techniques
Manuel Carro

IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2015

Predictive business process monitoring aims at forecasting potential problems during process execution before they occur so that these problems can be handled proactively. Several predictive monitoring techniques have been proposed in the past. However, so far those prediction techniques have been assessed only independently from each other, making it hard to reliably compare their applicability and accuracy. We empirically analyze and compare three main classes of predictive monitoring techniques, which are based on machine learning, constraint satisfaction, and Quality-of-Service (QoS) aggregation. Based on empirical evidence from an industrial case study in the area of transport and logistics, we assess those techniques with respect to five accuracy indicators. We further determine the dependency of accuracy on the point in time during process execution when a prediction is made in order to determine lead-times for accurate predictions. Our evidence suggests that, given a lead-time of half of the process duration, all predictive monitoring techniques consistently provide an accuracy of at least 70%. Yet, it also becomes evident that the techniques differ in terms of how accurately they may predict violations and nonviolations. To improve the prediction process, we thus exploit the characteristics of the individual techniques and propose their combination. Based on our case study data, evidence indicates that certain combinations of techniques may outperform individual techniques with respect to specific accuracy indicators. Combining constraint satisfaction with QoS aggregation, for instance, improves precision by 14%; combining machine learning with constraint satisfaction shows an improvement in recall by 23%.

View PDFchevron_right
Predictive monitoring of temporally-aggregated performance indicators of business processes against low-level streaming events
Massimo Guarascio

Information Systems, 2018

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Highlights • We focus on the context of complex event-based information system monitoring. • We provide a framework for supporting predictive monitoring of process performance indicators. • We face the problem of predicting whether (the process instances in) each current time window will violate an aggregate performance constraint, at different checkpoints inside the window.

View PDFchevron_right
Discovering Context-Aware Models for Predicting Business Process Performances
luigi pontieri

Lecture Notes in Computer Science, 2012

Discovering predictive performance models, offering run-time support, is an emerging topic in Process Mining research, which can effectively help the optimization of business process enactments. However, making accurate estimates is not easy especially when considering fine-grain performance measures (e.g., processing times) on a complex and flexible business process, where performance patterns change over time, depending on both case properties and context factors (e.g., seasonality, workload). We try to face such a situation by an ad-hoc predictive clustering approach, where different context-related execution scenarios are discovered and modeled accurately via distinct state-aware performance predictors. A readable predictive model is obtained eventually, which can make performance forecasts for any new running process case, by using the predictor of the cluster it is estimated to belong to. To ensure satisfactory levels of effectiveness and scalability, a concise encoding is used for each log trace, combining both context data and a series of values of the target measure, kept at certain key points in the trace. The approach was implemented in a system prototype, and validated on a real-life context. Test results confirmed the scalability of the approach, and its efficacy in predicting both processing times and associated SLA violations.

View PDFchevron_right
A test-bed for the evaluation of business process prediction techniques
suraj pandey

2011

Business process prediction technologies are being increasingly used by organisations to provide timely feedback to their customers and improve their overall productivity. In order to provide valuable information to both customers and system managers, the timings of business processes need to be forecast with high accuracy and efficiency. In particular, organisations require to predict the process and event flows, recognize their patterns, and forecast the total time it would take for a workflow to complete, in order to meet the Service Level Agreements (SLAs) signed with the customers. In this paper, we focus on the prediction models that could be used for forecasting time to completion of business processes by analysing historical event logs. First, we propose a service oriented architecture that provides a test-bed for carrying out predictions on business processes. Second, we propose a Hidden Markov Model (HMM) based prediction technique that produces a model based on event logs, and compare it against existing prediction models. Finally, we describe an implementation of the system, where we simulate the execution of a business process and obtain predictions using both the proposed and existing prediction techniques.

View PDFchevron_right
Monitoring and Analyzing Influential Factors of Business Process Performance
Florian Rosenberg

2009 IEEE International Enterprise Distributed Object Computing Conference, 2009

Business activity monitoring enables continuous observation of key performance indicators (KPIs). However, if things go wrong, a deeper analysis of process performance becomes necessary. Business analysts want to learn about the factors that influence the performance of business processes and most often contribute to the violation of KPI target values, and how they relate to each other. We provide a framework for performance monitoring and analysis of WS-BPEL processes, which consolidates process events and Quality of Service measurements. The framework uses machine learning techniques in order to construct tree structures, which represent the dependencies of a KPI on process and QoS metrics. These dependency trees allow business analysts to analyze how the process KPIs depend on lower-level process metrics and QoS characterisitics of the IT infrastructure. Deeper knowledge about the structure of dependencies can be gained by drill-down analysis of single factors of influence.

View PDFchevron_right
Context-Aware Prediction on Business Process Executions
Massimo Guarascio

2012

Discovering predictive performance models is an emerging topic in Process Mining. However, making accurate estimates is not easy especially when considering fine-grain metrics (such as processing times) on complex and flexible processes, where performances may change over time depending on context factors. We try to face such a situation by a general predictive-clustering approach, where different context-related execution scenarios are find and equipped with distinct performanceprediction models. A two-stage forecast can be then made for a new process case by using the model of the cluster it is estimated to belong to. Tests on real-life logs confirmed the validity of the approach.

View PDFchevron_right
Process Mining to Forecast the Future of Running Cases
Sonja Pravilovic

Lecture Notes in Computer Science, 2014

Processes are everywhere in our daily lives. More and more information about executions of processes are recorded in event logs by several information systems. Process mining techniques are used to analyze historic information hidden in event logs and to provide surprising insights for managers, system developers, auditors, and end users. While existing process mining techniques mainly analyze full process instances (cases), this paper extends the analysis to running cases, which have not yet completed. For running cases, process mining can be used to notify future events. This forecasting ability can provide insights for check conformance and support decision making. This paper details a process mining approach, which uses predictive clustering to equip an execution scenario with a prediction model. This model accounts for recent events of running cases to predict the properties of future events. Several tests with benchmark logs investigate the viability of the proposed approach.

View PDFchevron_right
Outcome-Oriented Predictive Process Monitoring
Marlon Dumas

ACM Transactions on Knowledge Discovery from Data

Predictive business process monitoring refers to the act of making predictions about the future state of ongoing cases of a business process, based on their incomplete execution traces and logs of historical (completed) traces. Motivated by the increasingly pervasive availability of fine-grained event data about business process executions, the problem of predictive process monitoring has received substantial attention in the past years. In particular, a considerable number of methods have been put forward to address the problem of outcome-oriented predictive process monitoring, which refers to classifying each ongoing case of a process according to a given set of possible categorical outcomes—e.g., Will the customer complain or not? Will an order be delivered, canceled, or withdrawn? Unfortunately, different authors have used different datasets, experimental settings, evaluation measures, and baselines to assess their proposals, resulting in poor comparability and an unclear pictur...

View PDFchevron_right
A Data-adaptive Trace Abstraction Approach to the Prediction of Business Process Performances
Luigi Pontieri

This paper presents a novel approach to the discovery of predictive process models, which are meant to support the run-time prediction of some performance indicator (e.g., the remaining processing time) on new ongoing process instances. To this purpose, we combine a series of data mining techniques (ranging from pattern mining, to non-parametric regression and to predictive clustering) with ad-hoc data transformation and abstraction mechanisms. As a result, a modular representation of the process is obtained, where different performancerelevant variants of it are provided with separate regression models, and discriminated on the basis of context information. Notably, the approach is capable to look at the given log traces at a proper level of abstraction, in a pretty automatic and transparent fashion, which reduces the need for heavy intervention by the analyst (which is, indeed, a major drawback of previous solutions in the literature). The approach has been validated on a real application scenario, with satisfactory results, in terms of both prediction accuracy and robustness.

View PDFchevron_right
Extracting accurate performance indicators from execution logs using process models
Ahmed Awad

2015 IEEE/ACS 12th International Conference of Computer Systems and Applications (AICCSA), 2015

Performance indicators (PIs) are defined to measure the progress towards the achievement of goals. The input to calculate these measures are the daily operations of an organization. Business processes are the means by which an organization operationalizes the achievement of its goals. So, it is logical to measure PIs based on business processes performance. Processes manifest their performance in execution logs where process and context data are recorded. In this paper, we introduce a new approach that uses information stored in process execution logs in order to extract more detailed and accurate performance measures. We show how the inclusion of information about activity lifecycles and process models can help achieve that. In addition, we evaluate our approach against similar approaches by using both real and synthetic logs.

View PDFchevron_right

References (51)

  1. Maggi Fabrizio Maria, Di Francescomarino Chiara, Dumas Mar- lon, Ghidini Chiara. Predictive monitoring of business processes. In: CAiSE:457-472. Springer; 2014.
  2. Leontjeva Anna, Conforti Raffaele, Di Francescomarino Chiara, Dumas Marlon, Maggi Fabrizio Maria. Complex Symbolic Sequence Encodings for Predictive Monitoring of Business Processes. In: BPM:297-313; 2015.
  3. Spoel Sjoerd, Keulen Maurice, Amrit Chintan. Process prediction in noisy data sets: a case study in a Dutch hospital. In: International Symposium on Data-Driven Process Discovery and Analysis:60- 83. Springer; 2012.
  4. Evermann Joerg, Rehse Jana-Rebecca, Fettke Peter. A Deep Learn- ing Approach for Predicting Process Behaviour at Runtime. In: Busi- ness Process Management Workshops -BPM 2016 International Workshops, Rio de Janeiro, Brazil, September 19, 2016, Revised Papers:327-338; 2016.
  5. Aalst Wil M. P., Schonenberg M. H., Song Minseok. Time prediction based on process mining. Information systems. 2011;36(2):450-475.
  6. Rogge-Solti Andreas, Weske Mathias. Prediction of remaining ser- vice execution time using stochastic Petri nets with arbitrary firing delays. In: ICSOC:389-403. Springer; 2013.
  7. Rogge-Solti Andreas, Weske Mathias. Prediction of business pro- cess durations using non-Markovian stochastic Petri nets. Informa- tion Systems. 2015;54:1-14.
  8. Dumas Marlon, La Rosa Marcello, Mendling Jan, Reijers Hajo A.. Fundamentals of Business Process Management, Second Edition. . Springer; 2018.
  9. Verenich Ilya, Nguyen Hoang, La Rosa Marcello, Dumas Marlon. White-box prediction of process performance indicators via flow analysis. In: Proceedings of the 2017 International Conference on Software and System Process, Paris, France, ICSSP 2017, July 5-7, 2017:85-94; 2017.
  10. Márquez-Chamorro A. E., Resinas M., Ruiz-Corts A.. Predictive monitoring of business processes: a survey. IEEE Transactions on Services Computing. 2017;PP(99):1-1.
  11. Pika Anastasiia, Aalst Wil M P, Fidge Colin J, Hofstede Arthur H. M., Wynn Moe T. Predicting deadline transgressions using event logs. In: BPM:211-216. Springer; 2012.
  12. Conforti Raffaele, Leoni Massimiliano, La Rosa Marcello, Aalst Wil M. P., Hofstede Arthur H. M.. A recommendation system for pre- dicting risks across multiple business process instances. Decision Support Systems. 2015;69:1-19.
  13. Metzger Andreas, Franklin Rod, Engel Yagil. Predictive mon- itoring of heterogeneous service-oriented business networks: The transport and logistics case. In: 2012 Annual SRII Global Conference:313-322. IEEE; 2012.
  14. Dongen Boudewijn F, Crooy Ronald A, Aalst Wil M P. Cycle time prediction: when will this case finally be finished?. In: CoopIS:319- 336. Springer; 2008.
  15. Polato Mirko, Sperduti Alessandro, Burattin Andrea, Leoni Massi- miliano. Data-aware remaining time prediction of business process instances. In: 2014 International Joint Conference on Neural Net- works, IJCNN 2014:816-823; 2014.
  16. Leoni Massimiliano, Aalst Wil M. P., Dees Marcus. A general pro- cess mining framework for correlating, predicting and cluster- ing dynamic behavior based on event logs. Information Systems. 2016;56:235-257.
  17. Senderovich Arik, Weidlich Matthias, Gal Avigdor, Mandelbaum Avishai. Queue mining for delay prediction in multi-class service processes. Information systems. 2015;53:278-295.
  18. Senderovich Arik, Di Francescomarino Chiara, Ghidini Chiara, Jorbina Kerwin, Maggi Fabrizio Maria. Intra and Inter-case Fea- tures in Predictive Process Monitoring: A Tale of Two Dimensions. In: Lecture Notes in Computer Science, vol. 10445: :306-323. Springer; 2017.
  19. Kikas Riivo, Dumas Marlon, Pfahl Dietmar. Using dynamic and con- textual features to predict issue lifetime in GitHub projects. In: Pro- ceedings of the 13th International Conference on Mining Software Repositories, MSR:291-302; 2016.
  20. Rees-Jones Mitch, Martin Matthew, Menzies Tim. Better Predic- tors for Issue Lifetime. CoRR. 2017;abs/1702.07735.
  21. Lakshmanan Geetika T, Shamsi Davood, Doganata Yurdaer N, Unuvar Merve, Khalaf Rania. A Markov prediction model for data- driven semi-structured business processes. Knowledge and Informa- tion Systems. 2015;42(1):97-126.
  22. Breuker Dominic, Matzner Martin, Delfmann Patrick, Becker Jörg. Comprehensible predictive models for business processes. MIS Quarterly. 2016;40(4):1009-1034.
  23. Tax Niek, Verenich Ilya, La Rosa Marcello, Dumas Marlon. Pre- dictive Business Process Monitoring with LSTM Neural Networks. In: Advanced Information Systems Engineering -29th Interna- tional Conference, CAiSE 2017, Essen, Germany, June 12-16, 2017, Proceedings:477-492; 2017.
  24. Tan P.N., Steinbach M., Karpatne A., Kumar V.. Introduction to Data Mining. What's New in Computer Science Series. Pearson Educa- tion; 2013.
  25. Polyvyanyy Artem, Smirnov Sergey, Weske Mathias. The Tricon- nected Abstraction of Process Models. In: Lecture Notes in Com- puter Science, vol. 5701: :229-244. Springer; 2009.
  26. Armas-Cervantes Abel, Baldan Paolo, Dumas Marlon, García- Bañuelos Luciano. Diagnosing behavioral differences between business process models: An approach based on event structures. Information systems. 2016;56:304-325.
  27. Aalst Wil M. P.. Process Discovery: Capturing the Invisible. IEEE Comp. Int. Mag.. 2010;5(1):28-41.
  28. Augusto Adriano, Conforti Raffaele, Dumas Marlon, et al. Auto- mated Discovery of Process Models from Event Logs: Review and Benchmark. CoRR. 2017;abs/1705.02288.
  29. Yang Yong, Dumas Marlon, García-Bañuelos Luciano, Polyvyanyy Artem, Zhang Liang. Generalized aggregate Quality of Service com- putation for composite services. Journal of Systems and Software. 2012;85(8):1818-1830.
  30. Teinemaa Irene, Dumas Marlon, Maggi Fabrizio Maria, Di Francescomarino Chiara. Predictive Business Process Monitor- ing with Structured and Unstructured Data. In: Business Process Management -14th International Conference, BPM 2016, Rio de Janeiro, Brazil, September 18-22, 2016:401-417; 2016.
  31. Mitchell Tom M.. Machine learning. McGraw-Hill; 1997.
  32. Fayyad Usama M., Piatetsky-Shapiro Gregory, Smyth Padhraic, Uthurusamy Ramasamy, eds. Advances in Knowledge Discovery and Data Mining. AAAI/MIT Press; 1996.
  33. Chen Tianqi, Guestrin Carlos. XGBoost: A Scalable Tree Boosting System. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Fran- cisco, CA, USA, August 13-17, 2016:785-794; 2016.
  34. Fowler B., Rajendiran M., Schroeder T., Bergh N., Flower A., Kang H.. Predicting patient revisits at the University of Virginia Health System Emergency Department. In: 2017 Systems and Information Engineering Design Symposium (SIEDS):253-258; 2017.
  35. Möller A, Ruhlmann-Kleider V, Leloup C, et al. Photometric classifi- cation of type Ia supernovae in the SuperNova Legacy Survey with supervised learning. Journal of Cosmology and Astroparticle Physics. 2016;2016(12):008.
  36. Torlay L., Perrone-Bertolotti Marcela, Thomas E., Baciu Monica. Machine learning-XGBoost analysis of language networks to clas- sify patients with epilepsy. Brain Informatics. 2017;4(3):159-169.
  37. Olson Randal S., La Cava William, Mustahsan Zairah, Varik Akshay, Moore Jason H.. Data-driven Advice for Applying Machine Learn- ing to Bioinformatics Problems. CoRR. 2017;abs/1708.05070.
  38. Hastie Trevor, Tibshirani Robert, Friedman Jerome H.. The elements of statistical learning: data mining, inference, and prediction, 2nd Edi- tion. . Springer; 2009.
  39. Urraca-Valle Ruben, Antoñanzas Javier, Antoñanzas-Torres Fer- nando, Pisón Francisco Javier. Estimation of Daily Global Hor- izontal Irradiation Using Extreme Gradient Boosting Machines. In: International Joint Conference SOCO'16-CISIS'16-ICEUTE'16 - San Sebastián, Spain, October 19th-21st, 2016, Proceedings:105- 113; 2016.
  40. Bergstra James, Bengio Yoshua. Random Search for Hyper- Parameter Optimization. Journal of Machine Learning Research. 2012;13:281-305.
  41. Augusto Adriano, Conforti Raffaele, Dumas Marlon, La Rosa Mar- cello, Bruno Giorgio. Automated Discovery of Structured Process Models: Discover Structured vs. Discover and Structure. In: Con- ceptual Modeling -35th International Conference, ER 2016:313- 329; 2016.
  42. Andrews Robert, Suriadi Suriadi, Wynn Moe, et al. Comparing static and dynamic aspects of patient flows via process model visualisa- tions. Preprint available at https://eprints.qut.edu.au/102848/. 2016;.
  43. Raudys Sarunas, Jain Anil K.. Small Sample Size Effects in Statisti- cal Pattern Recognition: Recommendations for Practitioners. IEEE Trans. Pattern Anal. Mach. Intell.. 1991;13(3):252-264.
  44. Teinemaa Irene, Dumas Marlon, La Rosa Marcello, Maggi Fab- rizio Maria. Outcome-Oriented Predictive Process Monitoring: Review and Benchmark. CoRR. 2017;abs/1707.06766.
  45. Ranzato Marc'Aurelio, Huang Fu Jie, Boureau Y-Lan, LeCun Yann. Unsupervised Learning of Invariant Feature Hierarchies with Appli- cations to Object Recognition. In: 2007 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007), 18-23 June 2007, Minneapolis, Minnesota, USA; 2007.
  46. Hyndman Rob J, Koehler Anne B. Another look at measures of fore- cast accuracy. International Journal of Forecasting. 2006;22(4):679- 688.
  47. Pedregosa F., Varoquaux G., Gramfort A., et al. Scikit-learn: Machine Learning in Python. Journal of Machine Learning Research. 2011;12:2825-2830.
  48. Fukunaga Keinosuke, Hayes Raymond R.. Effects of Sample Size in Classifier Design. IEEE Trans. Pattern Anal. Mach. Intell.. 1989;11(8):873-885.
  49. Tabachnick Barbara G. Using multivariate statistics: Sas Workbook. Addison-Wesley; 1996.
  50. Kull Meelis, Flach Peter A.. Reliability Maps: A Tool to Enhance Probability Estimates and Improve Classification Accuracy. In: Machine Learning and Knowledge Discovery in Databases -Euro- pean Conference, ECML PKDD 2014:18-33; 2014.
  51. How to cite this article: I. Verenich, H. Nguyen, M. Dumas, M. La Rosa, and A. ter Hofstede (2017), Predicting Process Performance: A White- Box Approach, Journal of Software: Evolution and Process, 2017;00:1-6.

Related papers

White-box prediction of process performance indicators via flow analysis
Marlon Dumas

Proceedings of the 2017 International Conference on Software and System Process - ICSSP 2017, 2017

Predictive business process monitoring methods exploit historical process execution logs to provide predictions about running instances of a process, which enable process workers and managers to preempt performance issues or compliance violations. A number of approaches have been proposed to predict quantitative process performance indicators, such as remaining cycle time, cost, or probability of deadline violation. However, these approaches adopt a black-box approach, insofar as they predict a single scalar value without decomposing this prediction into more elementary components. In this paper, we propose a white-box approach to predict performance indicators of running process instances. e key idea is to rst predict the performance indicator at the level of activities, and then to aggregate these predictions at the level of a process instance by means of ow analysis techniques. e paper speci cally develops this idea in the context of predicting the remaining cycle time of ongoing process instances. e proposed approach has been evaluated on four real-life event logs and compared against several baselines.

View PDFchevron_right
Predicting process performance: A white‐box approach based on process models
Marlon Dumas

Journal of software, 2019

Predictive business process monitoring methods exploit historical process execution logs to provide predictions about running instances of a process. These predictions enable process workers and managers to preempt performance issues or compliance violations. A number of approaches have been proposed to predict quantitative process performance indicators for running instances of a process, including remaining cycle time, cost, or probability of deadline violation. However, these approaches adopt a black-box approach, insofar as they predict a single scalar value without decomposing this prediction into more elementary components. In this paper, we propose a white-box approach to predict performance indicators of running process instances. The key idea is to first predict the performance indicator at the level of activities, and then to aggregate these predictions at the level of a process instance by means of flow analysis techniques. The paper develops this idea in the context of predicting the remaining cycle time of ongoing process instances. The proposed approach has been evaluated on real-life event logs and compared against several baselines.

View PDFchevron_right
Survey and Cross-benchmark Comparison of Remaining Time Prediction Methods in Business Process Monitoring
Marlon Dumas

ACM Transactions on Intelligent Systems and Technology

Predictive business process monitoring methods exploit historical process execution logs to generate predictions about running instances (called cases) of a business process, such as the prediction of the outcome, next activity, or remaining cycle time of a given process case. These insights could be used to support operational managers in taking remedial actions as business processes unfold, e.g., shifting resources from one case onto another to ensure the latter is completed on time. A number of methods to tackle the remaining cycle time prediction problem have been proposed in the literature. However, due to differences in their experimental setup, choice of datasets, evaluation measures, and baselines, the relative merits of each method remain unclear. This article presents a systematic literature review and taxonomy of methods for remaining time prediction in the context of business processes, as well as a cross-benchmark comparison of 16 such methods based on 17 real-life data...

View PDFchevron_right
Prediction of Business Process Execution Time
Walid Ben Fradj

Lecture notes in networks and systems 715, 2022

The use of immense amounts of data on the execution of applications based on business processes can make it possible, thanks to Process Mining, to detect trends. Indeed, human intelligence in decisionmaking is enriched by Machine Learning in order to avoid bottlenecks, improve efficiency and highlight potential process improvements. In this research article, we present a method (BPETPM) for predictive monitoring of business processes. This method allows to predict the execution time of a business process according to the path followed by the process instance. It predicts whether a process instance will run in time or late. We follow the CRISP-DM approach, known in Data Science, to carry out our method. The input data for learning is extracted from the event logs saving the execution traces of the workflow engine of a BPMS. We start by cleaning data, adding additional attributes, and encoding categorical variables. Then, at the modelling level, we test six classification algorithms : KNN, SVM(kernel=linear), SVM(kernel=rbf), Decision Tree, Random Forest and Logestic Regression. Then, using the BPETPM method, we create an intelligent process management system (iBPMS4PET). This system is applied to a process for managing incoming mail in the mutual health sector.

View PDFchevron_right
Data-Driven Process Performance Measurement and Prediction: A Process-Tree-Based Approach
Wil van der Aalst

To achieve operational excellence, a clear understanding of the core processes of a company is vital. Process mining enables companies to achieve this by distilling historical process knowledge based on recorded historical event data. Few techniques focus on the prediction of process performance after process redesign. This paper proposes a foundational framework for a data-driven business process redesign approach, allowing the user to investigate the impact of changes in the process, w.r.t. the overall process performance. The framework supports the prediction of future performance based on anticipated activity-level performance changes and control-flow changes. We have applied our approach to several real event logs, confirming our approach's applicability.

View PDFchevron_right

Related topics

  • Computer Science
    • 580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved