Computing and Information Systems - Research Publications

Permanent URI for this collection

Search Results

Now showing 1 - 10 of 115
  • Item
    Thumbnail Image
    Agent Miner: An Algorithm for Discovering Agent Systems from Event Data
    Tour, A ; Polyvyanyy, A ; Kalenkova, A ; Senderovich, A ; Di Francescomarino, C ; Burattin, A ; Janiesch, C ; Sadiq, S (Springer Nature Switzerland, 2023)
    Process discovery studies ways to use event data generated by business processes and recorded by IT systems to construct models that describe the processes. Existing discovery algorithms are predominantly concerned with constructing process models that represent the control flow of the processes. Agent system mining argues that business processes often emerge from interactions of autonomous agents and uses event data to construct models of the agents and their interactions. This paper presents and evaluates Agent Miner, an algorithm for discovering models of agents and their interactions from event data composing the system that has executed the processes which generated the input data. The conducted evaluation using our open-source implementation of Agent Miner and publicly available industrial datasets confirms that our algorithm can provide insights into the process participants and their interaction patterns and often discovers models that describe the business processes more faithfully than process models discovered using conventional process discovery algorithms.
  • Item
    Thumbnail Image
    Fast and accurate data-driven goal recognition using process mining techniques
    Su, Z ; Polyvyanyy, A ; Lipovetzky, N ; Sardina, S ; van Beest, N (ELSEVIER, 2023-10)
    The problem of goal recognition requests to automatically infer an accurate probability distribution over possible goals an autonomous agent is attempting to achieve in the environment. The state-of-the-art approaches for goal recognition operate under full knowledge of the environment and possible operations the agent can take. This knowledge, however, is often not available in real-world applications. Given historical observations of the agents' behaviors in the environment, we learn skill models that capture how the agents achieved the goals in the past. Next, given fresh observations of an agent, we infer their goals by diagnosing deviations between the observations and all the available skill models. We present a framework that serves as an outline for implementing such data-driven goal recognition systems and its instance system implemented using process mining techniques. The evaluations we conducted using our publicly available implementation confirm that the approach is well-defined, i.e., all system parameters impact its performance, has high accuracy over a wide range of synthetic and real-world domains, which is comparable with the more knowledge-demanding state-of-the-art approaches, and operates fast.
  • Item
    Thumbnail Image
    Learning When to Treat Business Processes: Prescriptive Process Monitoring with Causal Inference and Reinforcement Learning
    Bozorgi, ZD ; Dumas, M ; Rosa, ML ; Polyvyanyy, A ; Shoush, M ; Teinemaa, I ; Indulska, M ; Reinhartz-Berger, I ; Cetina, C ; Pastor, O (Springer Nature Switzerland, 2023)
    Increasing the success rate of a process, i.e. the percentage of cases that end in a positive outcome, is a recurrent process improvement goal. At runtime, there are often certain actions (a.k.a. treatments) that workers may execute to lift the probability that a case ends in a positive outcome. For example, in a loan origination process, a possible treatment is to issue multiple loan offers to increase the probability that the customer takes a loan. Each treatment has a cost. Thus, when defining policies for prescribing treatments to cases, managers need to consider the net gain of the treatments. Also, the effect of a treatment varies over time: treating a case earlier may be more effective than later in a case. This paper presents a prescriptive monitoring method that automates this decision-making task. The method combines causal inference and reinforcement learning to learn treatment policies that maximize the net gain. The method leverages a conformal prediction technique to speed up the convergence of the reinforcement learning mechanism by separating cases that are likely to end up in a positive or negative outcome, from uncertain cases. An evaluation on two real-life datasets shows that the proposed method outperforms a state-of-the-art baseline.
  • Item
    Thumbnail Image
    Prescriptive process monitoring based on causal effect estimation
    Bozorgi, ZD ; Teinemaa, I ; Dumas, M ; La Rosa, M ; Polyvyanyy, A (PERGAMON-ELSEVIER SCIENCE LTD, 2023-06)
    Prescriptive process monitoring methods seek to control the execution of a business process by triggering interventions, at runtime, to optimise one or more performance measure(s) such as cycle time or defect rate. Examples of interventions include, for example, using a premium shipping service to reduce cycle time in an order-to-cash process, or offering better loan conditions to increase the acceptance rate in a loan origination process. Each of these interventions comes with a cost. Thus, it is important to carefully select the set of cases to which an intervention is applied. The paper proposes a prescriptive process monitoring method that incorporates causal inference techniques to estimate the causal effect of triggering an intervention on each ongoing case of a process. Based on this estimate, the method triggers interventions according to a user-defined policy, taking into account the net gain of the interventions. The method is evaluated on four real-life data sets.
  • Item
    Thumbnail Image
    There and Back Again
    Barenholz, D ; Montali, M ; Polyvyanyy, A ; Reijers, HA ; Rivkin, A ; van der Werf, JMEM ; Gomes, L ; Lorenz, R (Springer Nature Switzerland, 2023)
    A process discovery algorithm aims to construct a model from data generated by historical system executions such that the model describes the system well. Consequently, one desired property of a process discovery algorithm is rediscoverability, which ensures that the algorithm can construct a model that is behaviorally equivalent to the original system. A system often simultaneously executes multiple processes that interact through object manipulations. This paper presents a framework for developing process discovery algorithms for constructing models that describe interacting processes based on typed Jackson Nets that use identifiers to refer to the objects they manipulate. Typed Jackson Nets enjoy the reconstructability property which states that the composition of the processes and the interactions of a decomposed typed Jackson Net yields a model that is bisimilar to the original system. We exploit this property to demonstrate that if a process discovery algorithm ensures rediscoverability, the system of interacting processes is rediscoverable.
  • Item
    Thumbnail Image
    Learning When to Treat Business Processes: Prescriptive Process Monitoring with Causal Inference and Reinforcement Learning
    Bozorgi, ZD ; Dumas, M ; La Rosa, M ; Polyvyanyy, A ; Shoush, M ; Teinemaa, I ( 2023-03-06)
    Increasing the success rate of a process, i.e. the percentage of cases that end in a positive outcome, is a recurrent process improvement goal. At runtime, there are often certain actions (a.k.a. treatments) that workers may execute to lift the probability that a case ends in a positive outcome. For example, in a loan origination process, a possible treatment is to issue multiple loan offers to increase the probability that the customer takes a loan. Each treatment has a cost. Thus, when defining policies for prescribing treatments to cases, managers need to consider the net gain of the treatments. Also, the effect of a treatment varies over time: treating a case earlier may be more effective than later in a case. This paper presents a prescriptive monitoring method that automates this decision-making task. The method combines causal inference and reinforcement learning to learn treatment policies that maximize the net gain. The method leverages a conformal prediction technique to speed up the convergence of the reinforcement learning mechanism by separating cases that are likely to end up in a positive or negative outcome, from uncertain cases. An evaluation on two real-life datasets shows that the proposed method outperforms a state-of-the-art baseline.
  • Item
    Thumbnail Image
    Process model forecasting and change exploration using time series analysis of event sequence data
    De Smedt, J ; Yeshchenko, A ; Polyvyanyy, A ; De Weerdt, J ; Mendling, J (ELSEVIER, 2023-05)
    Process analytics is a collection of data-driven techniques for, among others, making predictions for individual process instances or overall process models. At the instance level, various novel techniques have been recently devised, tackling analytical tasks such as next activity, remaining time, or outcome prediction. However, there is a notable void regarding predictions at the process model level. It is the ambition of this article to fill this gap. More specifically, we develop a technique to forecast the entire process model from historical event data. A forecasted model is a will-be process model representing a probable description of the overall process for a given period in the future. Such a forecast helps, for instance, to anticipate and prepare for the consequences of upcoming process drifts and emerging bottlenecks. Our technique builds on a representation of event data as multiple time series, each capturing the evolution of a behavioural aspect of the process model, such that corresponding time series forecasting techniques can be applied. Our implementation demonstrates the feasibility of process model forecasting using real-world event data. A user study using our Process Change Exploration tool confirms the usefulness and ease of use of the produced process model forecasts.
  • Item
    Thumbnail Image
    Stochastic-aware precision and recall measures for conformance checking in process mining
    Leemans, SJJ ; Polyvyanyy, A (Elsevier BV, 2023-05-01)
    Process mining studies ways to improve real-world processes using historical event data generated by IT systems that support business processes of organisations. Given an event log of an IT system, process discovery algorithms construct a process model representing the processes recorded in the log, while conformance checking techniques quantify how well the discovered model achieves this objective. State-of-the-art discovery and conformance techniques either completely ignore or consider but hide from the users information about the likelihood of process behaviour. That is, the vast majority of the existing process discovery algorithms construct non-stochastic aware process models. Consequently, few conformance checking techniques can assess how well such discovered models describe the relative likelihoods of traces recorded in the log or how well they represent the likelihood of future traces generated by the same system. Note that this is necessary to support process simulation, prediction and recommendation. Furthermore, stochastic information can provide business analysts with further actionable insights on frequent and rare conformance issues. This article presents precision and recall measures based on the notion of entropy of stochastic automata, which are capable of quantifying and, hence, differentiating, between frequent and rare deviations of an event log and a process model that is enriched with the information on the relative likelihoods of traces it describes. An evaluation over several real-world datasets that uses our open-source implementation of the measures demonstrates the feasibility of using our precision and recall measures in industrial settings. Finally, we propose a range of intuitive desired properties that stochastic precision and recall measures should possess, and study our and other existing stochastic-aware conformance measures with respect to these properties.
  • Item
    Thumbnail Image
    All that glitters is not gold: Four maturity stages of process discovery algorithms
    van der Werf, JMEM ; Polyvyanyy, A ; van Wensveen, BR ; Brinkhuis, M ; Reijers, HA (PERGAMON-ELSEVIER SCIENCE LTD, 2023-03)
    A process discovery algorithm aims to construct a process model that represents the real-world process stored in event data well; it is precise, generalizes the data correctly, and is simple. At the same time, it is reasonable to expect that better quality input event data should lead to constructed process models of better quality. However, existing process discovery algorithms omit the discussion of this relationship between the inputs and outputs and, as it turns out, often do not guarantee it. We demonstrate the latter claim using several quality measures for event data and discovered process models. Consequently, this paper requests for more rigor in the design of process discovery algorithms, including properties that relate the qualities of the inputs and outputs of these algorithms. We present four incremental maturity stages for process discovery algorithms, along with concrete guidelines for formulating relevant properties and experimental validation. We then use these stages to review several state of the art process discovery algorithms to confirm the need to reflect on how we perform algorithmic process discovery.
  • Item
    Thumbnail Image
    ProLift: Automated Discovery of Causal Treatment Rules From Event Logs (Extended Abstract)
    Bozorgi, ZD ; Kopõlov, A ; Dumas, M ; Rosa, ML ; Polyvyanyy, A ; Hassani, M ; Koschmider, A ; Comuzzi, M ; Maria Maggi, F ; Pufahl, L (CEUR, 2022-01-01)
    ProLift is a Web-based tool that uses causal machine learning, specifically uplift trees, to discover rules for optimizing business processes based on execution data (event logs). ProLift allows users to upload an event log, to specify case treatments and case outcomes, and to visualize treatment rules that increase the probability of positive case outcomes. The target audience of ProLift includes researchers and practitioners interested in leveraging causal machine learning for process improvement.