Business Process Event Logs Research Articles

Sampling business process event logs with guarantees

SummaryEvent log sampling has emerged as a key research focus in the field of process mining, aiming to enhance the efficiency of various process mining tasks, including model discovery, conformance checking, and process prediction. However, current log sampling techniques often fail to ensure high‐quality sample logs. This paper introduces a novel framework to support efficient event log sampling without compromising the quality of the sample log compared to the original one. The approach revolves around the consideration of directly‐follows relation (DFR) among business tasks as the fundamental behavior unit of an event log. By ensuring the DFR equivalence between the original and sample logs, the proposed technique addresses the challenge of sample log quality from the model discovery point of view. The framework is instantiated by seven distinct sampling strategies each has its own specialty and is fully implemented in the open‐source process mining tool platform ProM. To validate its effectiveness, we conducted a comprehensive experimental evaluation using 12 publicly available real‐life event logs against state‐of‐the‐art sampling techniques. The results clearly demonstrate that our technique significantly improves model discovery efficiency while upholding high quality of the discovered models.

Cross-Instance Regulatory Compliance Checking of Business Process Event Logs

Cross-Instance Regulatory Compliance Checking of Business Process Event Logs

A Systematic Review of Anomaly Detection for Business Process Event Logs

A Systematic Review of Anomaly Detection for Business Process Event Logs

Trace selection and allocation for log-move audit of business process directly-follows graphs

The auditing of business process execution may involve investigating selected process instances (traces) to identify the root causes of discrepancies. Discrepancies are unexpected business process behavior. Such discrepancies might reveal fraud, inefficiencies, or a model that does not reflect the process execution. One such discrepancy is consecutive activities in the event log that the business process model does not predict, called log-moves. In this work, we investigate strategies to select the minimum cost set of traces from the event log such that each log-move appears in at least one selected trace. A trace cost is any measure of effort required to audit it, such as the estimated required time. We also investigate how to distribute the selected traces among the auditors in order to balance the total costs allocated to them. We provide a mixed integer programming (MIP) formulation and a greedy algorithm for this problem, and empirically evaluate these solution approaches using two real-world datasets. The CPLEX solver finds the optimal solution of the MIP formulation for all instances in less than 75 min but presents exponential growth of execution time as we increase the number of auditors. On the other hand, the greedy heuristic has simple implementation, does not require an external solver, has polynomial time, and typically loses less than 15% of the solution quality for our datasets. The log-moves are obtained from the Directly-Follows Graph (DFG) of the business process model and event log; therefore, the model may be represented by any notation that allows the DFG extraction. For the particular case of business processes modeled as Process Trees, we provide an efficient algorithm for the DFG extraction. We also make no assumptions on how the business process model was built, either manually or by a process mining algorithm.

Discovering Structural Errors From Business Process Event Logs

Process mining aims at discovering behavioral knowledge of business processes from their event logs, which has received an increasing attention in the era of cloud computing and big data. Surprisingly, to date, discovering structural errors (e.g., deadlocks and lack of synchronization) from event logs has not been considered in state-of-the-art process mining techniques. Moreover, existing process discovery approaches cannot be directly applied to event logs of processes with structural errors due to erroneous event occurrences caused by unsynchronized activities. To address this problem, we first preprocess the event log to obtain two separate event logs that are used to discover deadlocks and lack of synchronization, respectively. Erroneous event occurrences caused by unsynchronized activities are discarded in the two processed event logs, from which our error mining algorithms can discover all process fragments involving structural errors, without the need to obtain the overall process first. We implement our approach in a ProM plugin and evaluate it on event logs of real-life business processes, the results of which demonstrate that our approach can effectively and efficiently discover deadlocks and lack of synchronization if event logs contain sufficient event sequences.

Leveraging a Heterogeneous Ensemble Learning for Outcome-Based Predictive Monitoring Using Business Process Event Logs

Outcome-based predictive process monitoring concerns predicting the outcome of a running process case using historical events stored as so-called process event logs. This prediction problem has been approached using different learning models in the literature. Ensemble learners have been shown to be particularly effective in outcome-based business process predictive monitoring, even when compared with learners exploiting complex deep learning architectures. However, the ensemble learners that have been used in the literature rely on weak base learners, such as decision trees. In this article, an advanced stacking ensemble technique for outcome-based predictive monitoring is introduced. The proposed stacking ensemble employs strong learners as base classifiers, i.e., other ensembles. More specifically, we consider stacking of random forests, extreme gradient boosting machines, and gradient boosting machines to train a process outcome prediction model. We evaluate the proposed approach using publicly available event logs. The results show that the proposed model is a promising approach for the outcome-based prediction task. We extensively compare the performance differences among the proposed methods and the base strong learners, using also statistical tests to prove the generalizability of the results obtained.

Pattern-Based Reconstruction of Anomalous Traces in Business Process Event Logs

Pattern-Based Reconstruction of Anomalous Traces in Business Process Event Logs

Leveraging an Advanced Heterogeneous Ensemble Learning for Outcome-Based Predictive Monitoring Using Business Process Event Logs

Leveraging an Advanced Heterogeneous Ensemble Learning for Outcome-Based Predictive Monitoring Using Business Process Event Logs

Trace Clustering: A Preprocessing Method to Improve the Performance of Process Discovery

The information system collects a large number of business process event logs, and process discovery aims to discover process models from the event logs. Many process discovery methods have been proposed, but most of them still have problems when processing event logs, such as low mining efficiency and poor process model quality. The trace clustering method allows to decompose original log to effectively solve these problems. There are many existing trace clustering methods, such as clustering based on vector space approaches, context-aware trace clustering, model-based sequence clustering, etc. The clustering effects obtained by different trace clustering methods are often different. Therefore, this paper proposes a preprocessing method to improve the performance of process discovery, called as trace clustering. Firstly, the event log is decomposed into a set of sub-logs by trace clustering method, Secondly, the sub-logs generate process models respectively by the process mining method. The experimental analysis on the datasets shows that the method proposed not only effectively improves the time performance of process discovery, but also improves the quality of the process model.

Https://ijsea.com/archive/volume10/issue9/IJSEA10091005.pdf

Massive amounts of business process event logs are collected and stored by modern information systems. Model discovery aims to discover a process model from such event logs, however, most of the existing approaches still suffer from low efficiency when facing large-scale event logs. Event log sampling techniques provide an effective scheme to improve the efficiency of process discovery, but the existing techniques still cannot guarantee the quality of model mining. Therefore, a sampling approach based on set coverage algorithm named set coverage sampling approach is proposed. The proposed sampling approach has been implemented in the open-source process mining toolkit ProM. Furthermore, experiments using a real event log data set from conformance checking and time performance analysis show that the proposed event log sampling approach can greatly improve the efficiency of log sampling on the premise of ensuring the quality of model mining.

Detecting anomalies in business process event logs using statistical leverage

The presence of anomalous information in a business process event log, such as missing, duplicated or swapped events, hampers the possibility of extracting useful insights from event log analysis. A number of approaches exist in the literature to detect anomalous cases in event logs based on different paradigms, such as probabilistic, distance-based or reconstruction-based anomaly detection. This paper proposes a novel method for anomaly detection in event logs based on the information-theoretic paradigm, which has not been considered before in event log anomaly detection. In particular, we propose an anomaly score for cases of a process based on statistical leverage and three different methods to set the anomaly detection threshold. The proposed approach does not require large data sets to train machine learning models, which are necessary for instance in reconstruction-based approaches. The proposed approach shows remarkable anomaly detection capability in experiments conducted using publicly available event logs in respect of existing methods in the literature. One of the proposed anomaly detection thresholds also shows to handle variable case anomaly ratios more effectively than other methods in the literature.

Sampling business process event logs using graph‐based ranking model

SummaryModern information systems are continuously collecting and storing large volumes of business process event logs. The analysis of event logs can provide valuable insights for business process re‐engineering and enhancement. Process discovery, as one of the most challenging event log analysis techniques, aims to discover a business process model from an event log. Many process discovery approaches have been proposed in the past two decades, however, most of them suffer from efficiency problem when dealing with large‐scale event logs. Motivated by PageRank, we propose LogRank, a graph‐based ranking model, for event log sampling in this paper. The LogRank is capable of sampling a large‐scale event log to a smaller size that can be efficiently handled by existing discovery approaches. To support real‐life applications, we instantiate the LogRank model for two typical types of event logs, that is, simple event logs and lifecycle event logs. To quantify the quality of a sample log with respect to the original one, we introduce a general evaluation framework that can be instantiated for different quality metrics. The proposed sampling approach has been implemented in the open‐source process mining toolkit ProM. By experiments with both synthetic and real‐life event logs, we demonstrate that the proposed LogRank‐based sampling approach provides an effective means to improve process discovery efficiency as well as guaranteeing high quality of discovered models.

An Empirical Investigation of Different Classifiers, Encoding, and Ensemble Schemes for Next Event Prediction Using Business Process Event Logs

There is a growing need for empirical benchmarks that support researchers and practitioners in selecting the best machine learning technique for given prediction tasks. In this article, we consider the next event prediction task in business process predictive monitoring, and we extend our previously published benchmark by studying the impact on the performance of different encoding windows and of using ensemble schemes. The choice of whether to use ensembles and which scheme to use often depends on the type of data and classification task. While there is a general understanding that ensembles perform well in predictive monitoring of business processes, next event prediction is a task for which no other benchmarks involving ensembles are available. The proposed benchmark helps researchers to select a high-performing individual classifier or ensemble scheme given the variability at the case level of the event log under consideration. Experimental results show that choosing an optimal number of events for feature encoding is challenging, resulting in the need to consider each event log individually when selecting an optimal value. Ensemble schemes improve the performance of low-performing classifiers in this task, such as SVM, whereas high-performing classifiers, such as tree-based classifiers, are not better off when ensemble schemes are considered.

Extraction of Missing Tendency Using Decision Tree Learning in Business Process Event Log

In recent years, process mining has been attracting attention as an effective method for improving business operations by analyzing event logs that record what is done in business processes. The event log may contain missing data due to technical or human error, and if the data are missing, the analysis results will be inadequate. Traditional methods mainly use prediction completion when there are missing values, but accurate completion is not always possible. In this paper, we propose a method for understanding the tendency of missing values in the event log using decision tree learning without supplementing the missing values. We conducted experiments using data from the incident management system and confirmed the effectiveness of our method.

BINet: Multi-perspective business process anomaly classification

In this paper, we introduce BINet, a neural network architecture for real-time multi-perspective anomaly detection in business process event logs. BINet is designed to handle both the control flow and the data perspective of a business process. Additionally, we propose a set of heuristics for setting the threshold of an anomaly detection algorithm automatically. We demonstrate that BINet can be used to detect anomalies in event logs not only at a case level but also at event attribute level. Finally, we demonstrate that a simple set of rules can be used to utilize the output of BINet for anomaly classification. We compare BINet to eight other state-of-the-art anomaly detection algorithms and evaluate their performance on an elaborate data corpus of 29 synthetic and 15 real-life event logs. BINet outperforms all other methods both on the synthetic as well as on the real-life datasets.

Autoencoders for improving quality of process event logs

Low quality of business process event logs, as determined by anomalous and missing values, is often unavoidable in practical contexts. The output of process analysis that uses event logs with missing and anomalous values is also likely to be of low quality, thus decreasing the quality of any decisions based on it. While previous work has focused on reconstructing missing events in an event log or removing anomalous traces, in this paper we focus on detecting anomalous values and reconstructing missing values at the level of attributes in event logs. We propose methods based on autoencoders, which are a class of neural networks that can reconstruct their own input and are particularly suitable to learn a model of the complex relationships among attribute values in an event log. These methods do not rely on any a-priori knowledge about the business process that generated an event log and are evaluated using real world and artificially-generated event logs. The paper also discusses a qualitative analysis of the impact of event log cleaning and reconstruction on the output of process discovery. The proposed approach shows remarkable performance regarding activity labels and timestamps in artificial event logs. The performance in the case of real world event logs, in particular timestamp anomaly detection, is lower, which may be due to high variability of attribute values in the chosen event logs. Process models discovered from reconstructed event logs are characterised by lower variability of allowed behaviour and, therefore, are more usable in practice.

An empirical comparison of classification techniques for next event prediction using business process event logs

Predictive analytics is an essential capability in business process management to forecast future status and performance of business processes. In this paper, we focus on one particular predictive monitoring task that is solved using classification techniques, i.e. predicting the next event in a case. Several different classifiers have been recently employed in the literature in this task. However, a quantitative benchmark of different classifiers is currently lacking. In this paper, we build such a benchmark by taking into account 20 classifiers from five families, i.e. trees, Bayesian, rule-based, neural and meta classifiers. We employ six real-world process event logs and consider two different sampling approaches, i.e. case and event-based sampling, and three different validation methods in order to acquire a comprehensive evaluation about the classifiers’ performance. According to our benchmark, the classifier most likely to be the overall superior performer is the credal decision tree (C-DT), followed by the other top-4 performers, i.e. random forest, decision tree, dagging ensemble, and nested dichotomies ensemble. We also provide a qualitative discussion of how features of an event log can affect the choice of best classifier.

Local Concurrency Detection in Business Process Event Logs

Process mining techniques aim at analyzing records generated during the execution of a business process in order to provide insights on the actual performance of the process. Detecting concurrency relations between events is a fundamental primitive underpinning a range of process mining techniques. Existing approaches to this problem identify concurrency relations at the level of event types under a global interpretation. If two event types are declared to be concurrent, every occurrence of one event type is deemed to be concurrent to one occurrence of the other. In practice, this interpretation is too coarse-grained and leads to over-generalization. This article proposes a finer-grained approach, whereby two event types may be deemed to be in a concurrency relation relative to one state of the process, but not relative to other states. In other words, the detected concurrency relation holds locally, relative to a set of states. Experimental results both with artificial and real-life logs show that the proposed local concurrency detection approach improves the accuracy of existing concurrency detection techniques.

Techniques of reordering traces in the event logs in business process management tasks

The problem of formation of input data for construction of models of business processes by methods of the intellectual analysis of processes is studied. It is shown that event logs of real information systems do not always meet the requirements of the ordering of input data on processes, that is, a log can have one track with events from different business processes. The tasks of analysis of trace characteristics in the business process event log and the development of methods for forming process events information order tracks organized on the basis of the comparison of event attributes and on the basis of comparison with the invariants of the event attributes of the business process logs are organized in the processes of the log of events events of the process management information system. The first method generates business process log lines based on pairwise comparison of event attributes based on the criterion for maximizing the number of shared attribute values for this pair. The method sequentially selects events from the common path for several business processes, and after pairwise comparison of the attributes with the events of the traces of specific business processes, determines the belonging of the next event to the path of the corresponding process. When comparing, the end event of the business process is also detected. The second method generates log lines based on the comparison of the invariant of the event attributes and the event, which is analyzed by the criterion of maximizing the weight of the values of the common attributes. Unlike the first method, when choosing a new business process, an invariant is created in the form of a sum of weights of the values of the attributes of the log of the business process. The scales of attribute values reflect the number of occurrences of these values in the execution of the business process. This allows you to take into account the history of the implementation of the business process in implementing the method. In practical terms, both proposed methods allow the formation of a set of business process event logs that are executed in parallel in a format suitable for use of methods and technologies of the intellectual analysis of proces ses. The first method has a lower accuracy. However, its advantage is the ability to use in the presence of only a joint path of several business processes, without the previously known orderly traces log of each business process. The second method allows you to increase the accuracy of the highlighting of the event path for each business process. The disadvantage of this method is that it requires a priori formation of the invariant of the attributes of the business process events.

Filtering Out Infrequent Behavior from Business Process Event Logs

In the era of “big data”, one of the key challenges is to analyze large amounts of data collected in meaningful and scalable ways. The field of process mining is concerned with the analysis of data that is of a particular nature, namely data that results from the execution of business processes. The analysis of such data can be negatively influenced by the presence of outliers, which reflect infrequent behavior or “noise”. In process discovery, where the objective is to automatically extract a process model from the data, this may result in rarely travelled pathways that clutter the process model. This paper presents an automated technique to the removal of infrequent behavior from event logs. The proposed technique is evaluated in detail and it is shown that its application in conjunction with certain existing process discovery algorithms significantly improves the quality of the discovered process models and that it scales well to large datasets.