Cross-System Process Mining using RFID Technology

Christian Glaschke, Norbert Gronau and Benedict Bender

Chair of Business Information Systems, especially Systems and Processes, University of Potsdam,

August-Bebel-Straße 89, Potsdam, Germany

{cglaschke, ngronau, bbender}@lswi.de

Keywords: Process Mining, RFID, Production, Cross-System.

Abstract: In times of digitalization, the collection and modeling of business processes is still a challenge for companies.

The demand for trustworthy process models that reflect the actual execution steps therefore increases. The

respective kinds of processes significantly determine both, business process analysis and the conception of

future target processes and they are the starting point for any kind of change initiatives. Existing approaches

to model as-is processes, like process mining, are exclusively focused on reconstruction. Therefore,

transactional protocols and limited data from a single application system are used. Heterogeneous application

landscapes and business processes that are executed across multiple application systems, on the contrary, are

one of the main challenges in process mining research. Using RFID technology is hence one approach to close

the existing gap between different application systems. This paper focuses on methods for data collection

from real world objects via RFID technology and possible combinations with application data (process

mining) in order to realize a cross system mining approach.

1 INTRODUCTION

Since more than 20 years, business process

management is the leading paradigm for organizing

and restructuring of corporations and public entities.

Although nearly all kinds of companies use business

process management in certain areas, there are still

some challenges that need further analysis. To name

only a few, we can draw the reader’s attention to the

following issues (Gronau, 2016):

• Improving learning while performing a business

process

• Improved usage of person-bound knowledge that

is generated in or used during the business

process.

• Establishment of PDCA cycles (plan-do-check-

act) in process management to be able to detect

deviations and to correct them without

interrupting the business process.

• Typical business processes nowadays are

supported by enterprise systems like ERP, CRM

or SCM systems. Nevertheless, there are normally

deviations between the intended process covered

by the ERP and the process actually performed by

the company (Gronau, 2015)

• Irrespective of the high degree of automation, in

most business processes, the human-machine

interface is more important than ever because

automated business processes inevitably have to

be interrupted to allow the human being to decide

or do something in order to propel the entire

process. Today, however, the description of

human interfaces is not intuitive at all.

• Last but not least, an increased real world

awareness for nowadays available business

process objects (like information, persons or

customers’ material) is needed. Also approaches

to integrate these kinds of information into the

respective processes are more necessary than

ever.

Valid models of business processes are the basis

for every optimization approach. Process mining has

already shown some merits in some realms and is able

to measure the actual performance of a business

process - partially with the help of existing process

models. Nevertheless, existing process mining

approaches are limited in their scope to a single

information system. Overcoming this barrier would

not only bring the monitoring closer to the actual

business process, it could also mean a big step ahead

in the direction of process model discovery.

A main obstacle on the road to model

improvement is the recognition of two entities with

179

Glaschke C., Gronau N. and Bender B.

Cross-System Process Mining using RFID Technology.

DOI: 10.5220/0006223501790186

In Proceedings of the Sixth International Symposium on Business Modeling and Software Design (BMSD 2016), pages 179-186

ISBN: 978-989-758-190-8

different IDs in different information systems being

one and the same (role, object, instance) in the real

business process. For this purpose, i.e. to close this

gap, this paper suggests to use RFID (Radio-

frequency identification) technology.

2 PROCESS MINING WITH RFID

The focus of the present paper lies on the recognition

of processes in heterogeneous application landscapes.

In the course of digitalization, a lot of data on a

company’s business processes have become

available. For instance, position vectors of real

objects by means of RFID technology (Zhang et al.,

2012), cyber-physical systems are able to realize

process components autonomously and application

systems have to be integrated over several levels in

order to meet the requirements. To develop a system

in this environment that allows for the reproduction

and understanding of the business process is therefore

the problem to be solved by the present work. This

problem has already been defined and treated in other

studies (Van der Aalst et al., 2012, Glaschke, 2015)

and it aims at overcoming the system barriers. To

achieve this aim, especially process mining is used,

as this approach allows for an identification of

process patterns through the analysis of a system’s

log files. The key idea of the present article is

therefore to combine process mining with the data

from RFID systems to thereby gain a comprehensive

view on business processes.

In order to solve this problem, this paper will

follow the Design Science Research Methodology

(DSRM). According to Peffers et al. (2007) the

DSRM involves six steps. Section one and two relate

to the first two steps of the DSRM approach, namely

the problem identification and definition of

objectives. Section three discusses related techniques

and technologies that could assist the cross-system

based process mining. Section four corresponds to the

design and development phase since it specifies the

solution for cross-system based process mining.

Furthermore, the second subsection demonstrates the

solution with a theoretical example which relates to

the demonstration phase. The evaluation of the

approach needs to be conducted in an experimental

setting for which the approach is described in the fifth

section of this paper. With regard to the

communication step the approach is presented in this

paper, while the evaluation results will be published

in further papers as well as necessary modifications

of the approach.

Finally, the results will be critically discussed and

an outlook will be given. The DSRM (Design Science

research methodology) seems appropriate since the

approach is motivated by theoretical as well as

practical issues and involves the design of an

appropriate solution for the given problem and it´s

evaluation.

3 RELATED WORK

For this paper, two research topics are of utmost

importance. First, we will focus on process mining to

identify the existing cross application approaches of

this method.

The second relevant research topic is the

technology of RFID. With respect to this issue, we

will particularly focus on works that link these two

approaches. Concerning the technology of RFID, we

will especially explore the question of how the data is

recorded in the application landscape and how it is

provided for further processing.

At the end of this section, a brief summary of the

existing research gap and the most important

corresponding findings will be given.

3.1 Process Mining

A well-known approach for process discovery is the

concept of process mining, which was developed by

Van der Aalst and his research group at the Technical

University of Eindhoven (The Netherlands). This

approach uses log files from application systems (for

instance ERP systems) to reconstruct process models.

To be successful in that effort, the application system

has to provide the information needed in a specific

manner (i.e. as shown in Table 1).

Table 1: Example of a log file.

PID Activity Worker Timestamp

452 registration 55 2011-12-24, 11:10:21

452 investigation 56 2011-12-24, 11:15:21

452 consulting 33 2011-12-24, 12:17:10

452 dismissal 55 2011-12-24, 12:47:11

453 registration 55 2011-12-24, 11:16:35

453 investigation 56 2011-12-24, 11:27:12

453 consulting 12 2011-12-24, 11:52:37

453 dismissal 55 2011-12-24, 11:59:54

454 registration 55 2011-12-24, 11:11:21

454 investigation 55 2011-12-24, 11:15:21

An important component needed for this listing of

Sixth International Symposium on Business Modeling and Software Design

180

process instances is the process identification number

(PID). This number is used to create a process

diagram based on more than just one process instance

(Van der Aalst, 2012). In the background, petri

networks are used here (i) to allow for the generation

of process diagrams, (ii) to describe the different

conditions of the process and (iii) to create a graph for

visualization purposes and analysis (Van der Aalst,

2011a; Accorsi, 2012).

In an article from Thiede and Fuerstenau (2016),

an extensive literature review on this topic is

presented. Therefore, the top 20 AIS Journals

between 2004 and 2015 are analysed. Additionally,

literature from production and organisational journals

have been screened for process mining and RFID

implementation publications. It could be concluded

that the usage of RFID to close the gap has not been

researched so far.

As can be deduced from Thiede and Fuerstenau

(2016), many research projects concentrate on the

field of “digital services” and among these especially

on the subtype of single systems. Thiede and

Fuerstenau (2016) define the concept of “service”

broadly as “the application of specialized knowledge

skills through deeds, processes, and performances for

the benefit of customers (Vargo and Lusch, 2004, p.

2). For instance, a customer buying an article in a web

shop (digital service) receives it by a dispatching

service (non-digital).”

Out of this studies the work on Cross-System

Process Mining is analyzed. There are different

approaches with wide scattered areas of application.

Now we describe the most important ones. The first

approach deals with data from smart environments.

The main result is an algorithm to mine frequent

patterns (Wen, 2015).

Another work deals with the theme of end to end

Process Management. The authors describe

implications for theory and practice based on the end

to end process management approach. Especially the

technological challenge is highlighted in the

investigation of end to end processes (Madder et al.,

2014).

Cross-organizational observations of processes,

however, have so far only been presented by very few

studies and also these work primarily focused the

aspect of workflow synchronization. An example of

this specific application of process mining is given by

Zeng et al., (2013).

Zeng et al., (2013) define in their work several

approaches for the synchronization of different log

files:

1. Coordinated with synchronized activities.

2. Coordinated with messages exchanged.

3. Coordinated with shared resources.

4. Coordinated with abstract procedures.

These four coordination patterns can generally be

used to link the different log files from various

companies. Nevertheless, this solution proposal is

based on the assumption that the respective processes

remain closed systems only linked to each other at

some specific points. As examples hereof, we can cite

the cases of goods delivery (activity) or document

exchange (messages). Based on these patterns, rules

can be defined that determine when exactly a

synchronization takes place in two log files or

models.

3.2 RFID

Today and due to digitalization, data on numerous

business objects is available. So, for instance, the

position vectors of an object can nowadays be

calculated in real-time due to RFID technologies

(Yingfeng et al., 2012). These data can then be used

to analyze the goods flow like demonstrated, for

example, in Jakkhupan (2012). For this purpose, the

objects involved in the process are first tracked and

then the respective data is saved.

Figure 1: RFID infrastructure (Krishna, 2007).

In a typical RFID infrastructure as shown in

Figure 1, the readers send unprocessed data to an

application system. In other architectures,

middleware has to be used for this purpose (Abad et

al., 2012). Furthermore, cross system communication

is made possible by writing collected information on

the RFID tag to be read out and further processed at

other points in the process (Finkenzeller, 2015).

A combination of both technologies, i.e. of RFID

and process mining, has already been used by Zhou

(2010) for the reconstruction of health care processes.

For this approach, the techniques of process mining

were applied to both, log files and the RFID

Cross-System Process Mining using RFID Technology

181

infrastructure. In doing so, it could be evidenced that

the log files of RFID infrastructures are also suitable

for being used in the context of process mining

applications and techniques (Gonzalez, 2006).

Fernandez et al., (2015) use a similar approach

working with indoor location systems to collect data

on different business objects in order to draw

conclusions concerning the corresponding processes.

For this second step, the authors also work with

process mining techniques.

3.3 Evaluation

Particularly from a reading of the work of Zeng et al.,

(2013), the question arises of how the present paper

will distinguish from previous research.

Synchronization of workflow models has already

been described extensively in literature. Hence, this

article primarily differs, on the one hand, with respect

to the investigated area, as here company internal

processes are investigated. On the other hand, in our

approach, the different workflow models from the

single application systems shall be integrated and not

just synchronized. Thus, we aim at establishing an

end-to-end business process that is interconnected

across the different system boundaries. To achieve

this goal, it is for example suitable to work with the

shared resources approach (Chapter 3.2), with the

respective data being tracked – in our case – by RFID

systems. In choosing this solution, it is not necessary

that both systems use the same denomination or ID

for one and the same object (Chapter 4).

4 CROSS SYSTEM PROCESS

MINING

In order to achieve the target to extract reliable and

encompassing models of process execution, it is

nowadays necessary to link information from

multiple information systems. This is due to the fact

that business processes involve different functional

areas within a company which are usually using

highly specialized software applications in order to

fulfill their tasks. As a result hereof, companies’

application landscapes involve many different

application systems, each executing a little part of the

overarching process (Fuerstenau and Glaschke,

2015).

4.1 Requirements and Objectives of the

Procedure

Established techniques like process mining are con-

ceptualized to evaluate information from single

information systems. Typically, for this purpose, log

files from one information system are analyzed. In

this procedure, firstly log entries are grouped

according to their affiliation (e.g. by assorting all log

entries that belong to a certain PID), secondly

dependencies and chronological consecutions are

analyzed and thirdly and finally all the

aforementioned bits of information are aggregated in

a workflow model.

In order to face now the challenge to model

complex cross-system processes, the concept of

process mining shall be complemented by a linking

technology in order to be able to model complete

business processes that are not limited to a single

information system. This linking technique shall

allow for a combination of the different sub-process

models that could be extracted from the different

information systems participating in the overarching

process.

One essential precondition of this approach is that

each information system that forms a part of the

overall process is connected with the RFID

infrastructure. This is necessary for the PID from the

information system to be passed to the RFID

infrastructure.

Another necessary prerequisite for this solution is

that the PIDs from the different information systems

are unique, i.e. that they are working with different

number ranges. Only in this way, each process

instance can be identified unequivocally.

Besides, also the given RFID tag must identify the

objects unequivocally by ignoring the sharing or

merging of RFID tag numbers.

Thus, our aim is to facilitate the integration of

various log files and the therein contained process

instances as illustrated in Figure 2.

Figure 2: Fundamental procedure for log file integration.

The combination of information system log files

Integration

Procedure

Logfile

ERP

Logfile

MES

Logfile

total

RFID

Data

Process Mining

Sixth International Symposium on Business Modeling and Software Design

182

with RFID data should reveal the overarching process

model which is composed of single system specific

process models that are combined in the appropriate

order. One of the major challenges in this context is

to identify corresponding sub-processes in the

different information systems. In order to face this

challenge, a distinct linkage between the different

sub-process models is needed.

Just like in case of process mining, configurations

have to be made within the information systems for

the created log files to provide an uninterrupted

description of the process instances. To achieve this,

each generated instance has to pass the current

process ID to the RFID infrastructure at known

interconnection points. This PID is then saved on the

respective RFID tag. To make this possible, scanning

processes must be performed at specific process

steps. Ideally, these scans will take place at the

beginning and end of a process.

In order to realize the linkage between the

different information systems, two kinds of physical

components are needed. The first component are

containers that physically accompany the different

sub-process steps through the entire process. This

kind of containers is set up in the first process, which

is executed in the ERP system. At this point, an empty

container is equipped with an RFID-tag that has a

unique ID.

The second component is the aforementioned

infrastructure of RFID gates, which is used to scan the

containers accompanying the different process steps.

Figure 3 illustrates how the individual scanning

processes are registered as entries on the RFID tag.

To allow for this kind of documentation, the involved

application systems have to be customized

accordingly. This customization procedure is already

standard in case of process mining applications (Van

der Aalst, 2011b). This technique, however,

additionally requires that when scanning the RFID

tags, the data is passed and stored. This also requires

the appropriate configuration of the participating

application systems.

Yet, this solution proposal is just one variant.

Another method would consist in passing the

unequivocal RFID tag ID to the information systems

to be saved there. In this case, the tag ID would have

to be written to the logs of the application systems.

Another possible approach would be to log the

identity within the RFID architecture. As here the

respective functionality needed to log the scan

processes is given, this variant will be favored in the

context of our validation.

Figure 3: Data structure on the RFID tag after scanning.

4.2 Approach for an Integration

Procedure

Although the here presented solution proposals seem

to be convincing, there is a significant gap between

the processing through information technology and

reality. Whereas in an information system tasks can

already be performed without the necessity of a

concrete processing object with RFID tag to exist, in

reality and as shown in figure 4, there will always be

overlapping time entries in the log files.

As strategy for our integration procedure, we

therefore have to take the following steps: The

process steps will have to be passed to the result log

following their chronological order and in doing so,

Figure 4: Basic procedure for log file integration (for one process instance).

ERP

Process ID - A212

Process ID - B315

MES

Process ID - C233

ID: 1223

RFID Tag

System Date

A212

B315

C233 MES

ERP

2015-10-24, 11:10:21

2015-10-26, 13:15:35

2015-10-28, 08:17:45

A212

ERP

2015-10-26, 09:08:43

scan

processes

ERP

Process ID - A212

Process flow

Step - 4

Step - 3

Step

Step - 24

Step - 30

PID

Logfiles

MES

ERP

Result

Cross-System Process Mining using RFID Technology

183

the PIDs from the application systems will have to be

replaced by the PID on the RFID tag. For this

purpose, the algorithm goes through the entries on the

RFID tag, systematically replacing the PIDs in the log

files, to afterwards copy the lines into one new

contiguous result log file.

The process mining procedure which further

processes the result log file then uses the

corresponding algorithms in order to reconstruct the

workflow with the aid of the time entries. Associated

examples and algorithms for this procedure are

presented by de Medeiros et al. (2003). For the

validation of this approach, an appropriate data model

is needed which is illustrated in Figure 5. Each RFID

reader is assigned to one place and one application

system. The RFID readers generate log entries,

which, in turn, are assigned to just one business

object. In our case, it is the read data and the written

data that is logged. Then, the software passes the

reading and writing processes to the assigned

application system.

Figure 5: Data Model of RFID Management Software.

Furthermore, we developed a process to be used

for validation that can also serve as reference for the

obtained result models. This process was designed to

reflect the typical production process of a

manufacturing company and it comprises the

following phases:

To begin with, one order with several items results

in different manufacturing jobs in the ERP system.

Next, the corresponding components are

commissioned and the production container is

packed. Afterwards, the order is scanned for the first

time and is passed on to the transport system (TS),

which moves the container to the first free

workstation. Subsequently, several production steps

are carried out in MES (Manufacturing Executin

System). At different process steps, the container is

scanned for identification purposes. In order to

measure the product properties, the data is handed

over to a quality management system and the

measurement results are documented. After this step,

the container is transported to the warehouse or

distribution center.

Even though the approach seems to solve a range

of problems with regard to the cross-system process

mining, it has to be remarked that this type of

aggregation cannot be used for all types of processes.

The applicability is limited to processes where the

physical parts, that are subsumed in the container, are

scanned at all of the involved information systems or

the data is passed, at least once, to each of the

participating information systems of the process. This

is due to the fact that the combination technique relies

on a briding mechnism based on part handover.

5 VALIDATION

In order to validate the concept for cross system

process modeling, several steps are necessary, which

need to be executed subsequently.

As a first step, an appropriate model composition

procedure needs to be realized in order to integrate

the different log files into one overarching log file. At

this point, and similar to the proceeding in process

model generation, the distinction of relevant and

irrelevant information determines the quality of the

composed overarching process model. This

composition process is supported by the information

from the RFID management software and, in

particular, by the unique IDs that are allocated by the

software component.

As a second step with regard to validation, the

experiment itself needs to be set up. Therefore, a

production setting is needed which allows for the

representation of the formerly defined process. This

stage involves physical components, such as some

kind of production machines, as well as the related

software infrastructure, which is composed of the

different systems mentioned in the process. Finally,

the RFID management software as well as the RFID

infrastructure including the containers have to be set

up.

For the validation of our approach, a university

Industry 4.0 Laboratory is used. This laboratory is

equipped with several application systems like,

among others, an ERP and a MES system from the

company Asseco Solutions AG. Additionally,

different infrastructure components like robots,

transportation systems and a RFID infrastructure are

available. The manufacturing process can thus be

simulated through configuration of the components or

it can be actually performed (figure 6).

The RFID infrastructure consists of stationary and

portable RFID readers with which RFID tag

information can be read and written. For the simple

configuration of the RFID infrastructure, we have a

Business

Object

Application

System

RFID reader

Log entry

Place

0..*

Sixth International Symposium on Business Modeling and Software Design

184

Figure 7: Example process with RFID support.

management software that is used as middleware in

the laboratory plant.

Figure 6: Physical validation environment.

Besides, delivery planning takes place while the

production is in progress. Based on the thereby

elaborated scheduling list, the TS decides where to

transport the products (warehouse vs. distribution

center). These steps are also logged and get assigned

the PID from the ERP (Enterprise Ressource

Planning) system.

For validation the process should be executed

several times with different variations. Especially the

different mining algorithms from the process mining

software (Claes and Poels, 2012) will here have to be

tested and the respective results will have to be

compared. The validation will be successful when the

elaborated models actually reflect the process shown

in Figure 7.

6 CONCLUSIONS AND

OUTLOOK

In this paper, it could be demonstrated that a simple

and accessible method can close the gap between the

different systems and processes. For this purpose, the

concept of shared resources (Zeng et al., 2013) was

applied, which is already used in cross-organizational

approaches. Furthermore, it could be shown that the

RFID technology does already provide a solution

concept. Thus, we have succeeded in presenting a

cross system integration procedure.

If the results are achieved within the laboratory

environment, the question arises of whether the

approach will also function in practice and in more

complex environments with massive parallel

operations. All in all, the question remains of what

factors are responsible for the quality of results. Here,

on the one hand, the quality of the log files from the

application systems will definitely have to be taken

into consideration and, on the other hand, the number

of iterations needed to identify parallel and

independent process steps will be of importance. The

validation results will be published in further works.

In this paper, a simple setting was chosen to

demonstrate the benefits of the technology

combination of process mining and RFID. Future

work must continue to follow this trend. It is possible

to track more than one and different business objects.

The application system provides further information,

which can be used for the enrichment of process

model analysis. An example - in this case products

were anonymous, but they can be categorized in

future according to the product category. So it is

possible to structure the mined process model.

To follow the idea of end to end processes we

have to consider the supplier processes and other

support processes like maintenance. Therefore

laboratory environments with real applications

systems and physical components can help, to

establish the techniques in practice by showing

practical solutions and related benefits.

REFERENCES

Abad, I., Cerrada, C., Cerrada, J. A., Heradio, R., Valero,

Cross-System Process Mining using RFID Technology

185

E., 2012. Managing RFID sensors networks with a

general purpose RFID middleware. Sensors, 12(6),

7719–7737.

Accorsi R., Stocker T., 2012. On the Exploitation of

Process Mining for Security Audits: The Conformance

Checking Case. ACM Symposium on Applied

Computing. doi:10.1145/2245276.2232051.

Claes, J., Poels, G., 2012. Process mining and the ProM

framework: An exploratory survey. Business Process

Management Workshops. Springer Berlin Heidelberg.

De Medeiros, A. A. K., Van Dongen, B F, Van der Aalst,

WMP., Weijters, A.J.M.M., 2003. Process mining:

extending the α-algorithm to mine short loops. On the

move to meaningful internet systems, Springer Berlin

Heidelberg, 389-406.

Fernandez-Latas, C., Lizondo, A., Monton, E., Benedi, J.

M., Traver, V., 2015. Process Mining Methodology for

Health Process Tracking Using Real-Time Indoor

Location Systems. Sensors, 15(12), 29821-29840.

Finkenzeller, K., 2015. RFID-Handbuch: RFID Handbook:

Fundamentals and practical applications of

transponders, contactless smart cards and NFC. (in

german) Carl Hanser Verlag, Volume 7.

Fuerstenau, D., Glaschke, C., 2015. Weighting of

Integration Qualities in IS Architectures: A Production

Case. Paper presented at the ECIS 2015, Germany.

Glaschke, C., Gronau, N., 2015. Automated Process Model

Discovery - Limitations and Challenges. Paper

presented at the International Symposium on Business

Modeling and Software Design.

Gonzalez, H., Han, J., Li, X., 2006. Mining Compressed

Commodity Workflows from Massive RFID Data Sets.

15th ACM International Conference on Information

and Knowledge Management, ACM, 162-171.

Gronau, N., 2016. Business Process Management in

industry and administration. (in german), Gito Verlag,

Berlin, 2016.

Gronau, N., 2015. Trends and Future Research in

Enterprise Systems. Lecture Notes in Business

Information Processing, Volume 198, 271-280.

Jakkhupan, W., Archint, S., Li, Y., 2012. Business process

analysis and simulation for the RFID and EPCglobal

Network enabled supply chain: A proof-of-concept

approach. Network and Computer Applications, 34(3),

949-957.

Krishna, P., Husalc, D., 2007. RFID Infrastructure. IEEE

Communications Magazine, 45(9), 4-10

Maddern, H., Smart, P. A., Maull, R. S., Childe, S., 2014.

End-to-End Process Management: Implications for

Theory and Practice. Production Planning & Control,

25(16), 1303-21.

Peffers, K., Tuunanen, T., Rothenberger, M., Chatterjee, S.,

2007. A Design Science Research Methodology for

Information Systems Research. Journal of Manage-

ment Information Systems. 24(3), 45-77.

Thiede, M., Fuerstenau, D., 2016. The Technological

Maturity of Process Mining: An Exploration of the

Status Quo in Top IS Journals. Paper presented at the

MKWI 2016, Germany.

Van der Aalst, WMP., Accorsi R., Ullrich M., 2012.

Process Mining. Retrieved. (in german) April 29, 2015,

from http://www.gi.de/nc/service/informatiklexikon

/detailansicht/article/process-mining.htm

Van der Aalst, W., 2011a. Process Mining - Discovery,

Conformance and Enhancement of Business Processes.

Springer.

Van der Aalst, W., 2011b. Service Mining: Using Process

Mining to Discover, Check, and Improve Service

Behavior. IEEE Transaction on Service Computing,

6(4), 525-535.

Vargo, S., Lusch, R., 2004. Evolving to a New Dominant

Logic for Marketing. Journal of Marketing, 68(1), 1-17.

Wen, J., Mingyang Z., Zhiying W., 2015. Activity

Recognition with Weighted Frequent Patterns Mining

in Smart Environments. Expert Systems with

Applications, 42, 6423-32.

Yingfeng, Z., Jiang, P., Huang, G. Qu, T., Zhou, G., Hong,

J., 2012. RFID-Enabled Real-Time Manufacturing

Information Tracking Infrastructure for Extended

Enterprises. Journal of Intelligent Manufacturing, 23,

2357-2366.

Zeng, Q., Sun, S. X., Duan, H., Liu, C., Wang, H., 2013.

Cross-Organizational Collaborative Workflow Mining

from a Multi-Source Log. Decision Support Systems

54, 1280-301.

Zhou, W., Piramuthu, S., 2010. Framework, strategy and

evaluation of health care processes with RFID.

Decision Support Systems, 50(1), 222-233.

Sixth International Symposium on Business Modeling and Software Design

186