Fire now, fire later: alarm-based systems for prescriptive process monitoring (2019-05-23T00:00:00.000000Z)

TL;DR

A framework for prescriptive process monitoring is proposed, which extends predictive monitoring with the ability to generate alarms that trigger interventions to prevent an undesired outcome or mitigate its effect and incorporates a parameterized cost model to assess the cost–benefit trade-off of generating alarms.

Abstract

Predictive process monitoring is a family of techniques to analyze events produced during the execution of a business process in order to predict the future state or the final outcome of running process instances. Existing techniques in this field are able to predict, at each step of a process instance, the likelihood that it will lead to an undesired outcome. These techniques, however, focus on generating predictions and do not prescribe when and how process workers should intervene to decrease the cost of undesired outcomes. This paper proposes a framework for prescriptive process monitoring, which extends predictive monitoring with the ability to generate alarms that trigger interventions to prevent an undesired outcome or mitigate its effect. The framework incorporates a parameterized cost model to assess the cost–benefit trade-off of generating alarms. We show how to optimize the generation of alarms given an event log of past process executions and a set of cost model parameters. The proposed approaches are empirically evaluated using a range of real-life event logs. The experimental results show that the net cost of undesired outcomes can be minimized by changing the threshold for generating alarms, as the process instance progresses. Moreover, introducing delays for triggering alarms, instead of triggering them as soon as the probability of an undesired outcome exceeds a threshold, leads to lower net costs.

Authors

Niek Tax

3 papers

M. Dumas

4 papers

Irene Teinemaa

3 papers

TL;DR

Abstract

Authors

References52 items

Time Matters: Time-Aware LSTMs for Predictive Business Process Monitoring

An ensemble-based approach to the security-oriented classification of low-level log traces

Prescriptive Business Process Monitoring for Recommending Next Best Actions

Extraction, correlation, and abstraction of event data for process mining

Information-preserving abstractions of event data in process mining

Proactive Process Adaptation Using Deep Learning Ensembles

Early Classification of Time Series by Simultaneously Optimizing the Accuracy and Earliness

Split miner: automated discovery of accurate and simple business process models from event logs

Alarm-Based Prescriptive Process Monitoring

Efficiently interpreting traces of low level events in business process logs

Enhancing Process Models to Improve Business Performance: A Methodology and Case Studies

Data-driven advice for applying machine learning to bioinformatics problems

Outcome-Oriented Predictive Process Monitoring

Predictive Business Process Monitoring Considering Reliability Estimates

Event stream-based process discovery using abstract representations

Cost-Aware Early Classification of Time Series

Prescriptive Control of Business Processes

Reliable early classification of time series based on discriminating the classes over time

Process Mining

A general process mining framework for correlating, predicting and clustering dynamic behavior based on event logs

Early Classification of Time Series as a Non Myopic Sequential Decision Making Problem

Clustering-Based Predictive Process Monitoring

Comparing and Combining Predictive Business Process Monitoring Techniques

A recommendation system for predicting risks across multiple business process instances

Prescriptive Analytics for Recommendation-Based Business Process Optimization

Predictive Monitoring of Business Processes

Algorithms for Hyper-Parameter Optimization

Early classification on time series

Multi-Label Classification: An Overview

Thresholding for Making Classifiers Cost-sensitive

Sequential cost-sensitive decision making with reinforcement learning

Reducing multiclass to binary: a unifying approach for margin classifiers

Learning and making decisions when costs and probabilities are both unknown

The Foundations of Cost-Sensitive Learning

Reducing Multiclass to Binary: A Unifying Approach for Margin Classifiers

MetaCost: a general method for making classifiers cost-sensitive

The Cost of Compensation

Unemployment

Do we need hundreds of classifiers to solve real world classification problems?

Classifying with confidence from incomplete information

Process Mining

Probabilistic Outputs for Support vector Machines and Comparisons to Regularized Likelihood Methods

Single-alarm cost model) A single-alarm cost model is a tuple ( c in , c out , c com , consisting of: alarm model ( c in , c com , eff )

log contains traces of a loan application process in a Dutch bank. 1 It was split into two sub-logs, denoted with bpic2017_refused and bpic2017_cancelled

Cost of intervention The card is automatically blocked by the system

Cost of intervention For the intervention, an employee needs to check if the customer is indeed receiving unentitled beneﬁts and, if so, ﬁll in the forms for stopping the payments

Cost of undesired outcome The amount of unentitled beneﬁts that the customer would obtain without stopping the payments

Fire now, ﬁre later.

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations

Fahrenkrog-Petersen is a final-year PhD student, working on Process Mining, an emerging subfield of Data Mining

Road trafﬁc ﬁnes This log originates from an Italian police unit and relates to a process to collect trafﬁc ﬁnes. 2 The desired outcome is that a ﬁne is paid, while in the

RQ1 How good are the thresholds identiﬁed by empirical thresholding in terms of the reduction of the average processing cost for different alarm model conﬁgurations?

Field of Study

Journal Information

Name

Page

Volume

Venue Information

Name

Type

URL

Alternate Names