Approaching Process Mining with Sequence Clustering: Experiments and Findings

2012

Information Systems

385

272

Performing business process analysis in healthcare organizations is particularly difficult due to the highlydynamic, complex, ad-hoc, and multi-disciplinary nature of healthcare processes. Process mining is a promising approach to obtain a better understanding about those processes by analyzing event data recorded in healthcare information systems. However, not all process mining techniques perform well in capturing the complex and ad-hoc nature of clinical workflows. In this work we introduce a methodology for the application of process mining techniques that leads to the identification of regular behavior, process variants, and exceptional medical cases. The approach is demonstrated in a case study conducted at a hospital emergency service. For this purpose, we implemented the methodology in a tool that integrates the main stages of process analysis. The tool is specific to the case study, but the same methodology can be used in other healthcare environments.

Section: Discussionmentioning

confidence: 99%

“…The Sequence Clustering plug-in [37] was motivated by previous work outside the ProM framework [38,39,40]. Sequence clustering takes the sequences of tasks that describe each trace and groups similar sequences into the same cluster.…”

Section: Clustering Techniquesmentioning

confidence: 99%

Business process analysis in healthcare environments: A methodology based on process mining

Rebuge

2012

Information Systems

385

272

“…The sequences of context changes for all users were then clustered together, without user identification, in order to check that it would be possible to rediscover the different user profiles. The dataset was stored in a Microsoft SQL Server database and then analyzed with SQL Server Analysis Services, in particular the sequence clustering algorithm which has been used in the past by our group to conduct other case studies as well [24]. It should be noted that users are free to define their own context names and therefore, in general, there will be a small but varying degree of overlap between the contexts defined by different users.…”

Section: Context Publicationmentioning

confidence: 99%

“…Our system automates this status updates by publishing the user context periodically, so that the user profile is always up-to-date. Context information is published over a Wi-Fi or GPRS/3G connection, whichever is available at the moment, using the official Hi5 REST API 24 . Table 5 presents the time it takes to publish to the Hi5 servers.…”

Section: Twitter and Hi5mentioning

confidence: 99%

Providing user context for mobile and social networking applications

Santos

Cardoso

Pervasive and Mobile Computing

et al. 2010

The processing capabilities of mobile devices coupled with portable and wearable sensors provide the basis for new context-aware services and applications tailored to the user environment and its daily activities. In this article, we describe the approach developed within the UPCASE project, which makes use of sensors available in the mobile device as well as sensors externally connected via Bluetooth to provide user contexts. We describe the system architecture from sensor data acquisition to feature extraction, context inference and the publication of context information in web-centered servers that support well-known social networking services. In the current prototype, context inference is based on decision trees to learn and to identify contexts dynamically at run-time, but the middleware allows the integration of different inference engines if necessary. Experimental results in a real-world setting suggest that the proposed solution is a promising approach to provide user context to local mobile applications as well as to network-level applications such as social networking services.

“…But if we look at bottomup approaches, and in particular to the problem of discovering business processes from event logs, we realize that current process mining techniques are able to recover process models in a representation that might not always match what the end users may be familiar with. Some techniques generate dependency graphs [15,16], others use probabilistic models [17,18], and most of the current techniques are geared towards retrieving Petri net models [19]. To communicate with end users, it may be necessary to translate Petri nets into other kinds of models, including business process notations such as EPC 6 [20,21].…”

Section: Introductionmentioning

confidence: 99%

Automatic Extraction of Process Control Flow from I/O Operations

Diniz

Lecture Notes in Computer Science

2008

Abstract. Many end users will expect the output of process mining to be a model they can easily understand. On the other hand, knowing which objects were accessed in each operation can be a valuable input for process discovery. From these two trends it is possible to establish an analogy between process mining and the discovery of program structure. In this paper we present an approach for extracting process controlflow from a trace of read and write operations over a set of objects. The approach is divided in two independent phases. In the first phase, Fourier analysis is used to identify periodic behavior that can be represented with loop constructs. In the second phase, a match-and-merge technique is used to produce a control-flow graph capable of generating the input trace and thus representing the process that generated it. The combination of these techniques provides a structured and compact representation of the unknown process, with very good results in terms of conformance metrics.