Applying clustering and classification data mining techniques for competitive and knowledge‐intensive processes improvement

• Date: 01 April 2019
• Author: Farzad Movahedi Sobhani,Mohammad Khanbabaei,Mahmood Alborzi,Reza Radfar
• DOI: http://doi.org/10.1002/kpm.1595
• Published date: 01 April 2019

RESEARCH ARTICLE

Applying clustering and classification data mining techniques

for competitive and knowledge‐intensive processes

improvement

Mohammad Khanbabaei

1

|Mahmood Alborzi

1

|Farzad Movahedi Sobhani

2

|Reza Radfar

3

1

Department of Information Technology

Management, Science and Research Branch,

Islamic Azad University, Tehran, Iran

2

Department of Industrial Engineering,

Science and Research Branch, Islamic Azad

University, Tehran, Iran

3

Department of Technology Management,

Science and Research Branch, Islamic Azad

University, Tehran, Iran

Correspondence

Mohammad Khanbabaei, Department of

Information Technology Management, Science

and Research Branch, Islamic Azad University,

Tehran, Iran.

Email: mohammadkhanbabaei@srbiau.ac.ir

Processes as one of the valuable knowledge resources can create sustainable compet-

itive advantages in organizations. There is a large number of processes in organiza-

tions. They generate a high volume of process data that leads to the high‐

dimensionality problems, complex relationships, dynamic changes, and difficulties in

the understanding of the process by human resources. Traditional process improve-

ment methodologies have weaknesses in environment with the large number of pro-

cesses. Data mining techniques can support process improvement in this

environment. They can recommend the improvement suggestions through extracting

valuable patterns from a high volume of the process dataset. Recently, knowledge‐

intensive processes have been increasingly concentrated in the field of process

improvement. These types of processes can induce a competitive behavior over the

other processes. The main problem is the improvement of competitive and

knowledge‐intensive processes in a high volume of process dataset.

The main purpose of this paper is to present a model to identify the behavior of compet-

itive and knowledge‐intensive processes and recommend improvement suggestions.

For this purpose, data mining techniques are applied to extract valuable patterns hidden

in a high volume of process dataset. In this regard, K‐means clustering and C5 classifi-

cation algorithms are applied to extract valuable patterns. A real process dataset was

used to evaluate the effectiveness and applicability of the model. The results confirmed

that the proposed model can apply data mining techniques to support competitive and

knowledge‐intensive process improvement in a high volume of process dataset.

KEYWORDS

Data mining, process improvement, competitive process, knowledge‐intensive process

1|INTRODUCTION

Processes as one of the most critical resources in organizations can

create competitive advantages. Gómez‐Pérez, Erdmann, Greaves,

Corcho, and Benjamins (2010) expressed that processes are the

operations that produce outputs by using the resources such as

event, time, place, expertise, and so forth. There is a high volume of

diverse processes in different organizations. Chen and Wang (1999)

presented characteristics of processes including the following: high

volumes and high dimensionality of the process dataset; noise and

uncertainty related to processes; dynamic and complex behavior of

the process features; and various measures for evaluating the

processes.

In addition, Goedertier (2008) stated that increasing the scale and

complexity of processes and their interrelationships is one of the

major challenges of organizations. Rebuge and Ferreira (2012) clarified

some other behaviors of the processes in organizations, including

dynamism and changes in processes, existence of numerous effective

Received: 8 September 2017 Accepted: 4 December 2018

DOI: 10.1002/kpm.1595

Knowl Process Manag. 2019;26:123–139. © 2019 John Wiley & Sons, Ltd.wileyonlinelibrary.com/journal/kpm 123

factors on processes, high volumes of the process dataset, and the

relationships of processes with different departments.

Process improvement is one of the main tasks in organizations.

However, there are some problems in process improvement, when

organization encounters with a high volume of the process dataset.

Jeong, Song, Shin, and Rae Cho (2008) numerated some of them as

follows: large number of processes along with their numerous fea-

tures; complexity and high‐dimensionality problems; difficulties in

finding hidden and suitable patterns embedded in the process dataset;

and time‐consuming and costly application of the process improve-

ment methodologies. Moreover, a single and isolated view to pro-

cesses is a dominant viewpoint in the vast majority of the process

improvement methodologies (Houy, Fettke, Loos, van der Aalst, &

Krogstie, 2011) and (Wuest, Irgens, & Thoben, 2013).

In addition, there is no comprehensive view to intelligent and

information‐basedp rocess improvement methodologies(Del gado, Weber,

Ruiz, De Guzmán, & Piattini, 2014). Past processimprovement methodol-

ogies do not have a respectable consideration to the interrelationships and

effects between processes (Lepmets, McBride, & Ras, 2012). There is a

rigidity in gathering, updating, and accessing the process dataset in process

improvement methods (Geffen & Niks, 2013). Furthermore, one of the

most important problems in the process improvement methodologies

refers to the process feature selection (Darmani & Hanafizadeh, 2013).

In this regard, Zelt, Schmiedel, and vom Brocke (2018) stated that previous

process management approaches considered only a few characteristics

to identify the behavior of processes. Although, it is necessary to

identify several features for evaluating the behavior of processes.

With respectto the abovementionedissues, data mining techniques

can support processimprovement by extractingvaluable patterns in the

large number of the processes and recommend improvement sugges-

tions basedon these patterns. Jung,Choi, and Song (2007) declaredthat

the information ofprocesses can be translated and extractedas a valu-

able knowledge for organizations. However, there is a low attention to

processes asknowledge (Lodhi, Köppen, & Saake,2013).

Some studies applieddata mining for processimprovement,as elab-

oratedin Section3.1. However,Jeong et al. (2008)explained thatbecause

there is a little attention to data mining in process improvement, these

studies have been accompanied with some weaknesses.Rupnik and Jaklic

(2009) declared that thereare some gaps for achievinga suitable level in

the application of data mining for the improvement of processes. In

addition, they stated that the previous studies used data mining for pro-

cess reengineering, in only the static and formal procedures. Moreover,

Wegener and Rüping (2010) indicated that there is a notable lack in

framework of using data mining solutions for process management.

In the other direction, traditional process improvement methodol-

ogies are often process‐oriented (Seethamraju & Marjanovic, 2009).

They do not properly consider the notion of knowledge in their pro-

cesses (Davenport, 2010). El Sawy and Josefek (2004) confirmed that

the next wave of process improvement is based on knowledge. More-

over, several studies have considered the combination of knowledge

management approach and process improvement. For example,

Massingham and Al Holaibi (2017) presented how knowledge manage-

ment can be used for improving processes. On the other side, the pro-

cess improvement methodologies have newly focused on knowledge‐

intensive processes (Ciccio, Marrella, & Russo, 2015).

Marjanovic and Freeze (2012) declared that knowledge‐intensive

processes can lead to competitive differentiation. These processes can

be regarded as competitive processes in organizations. They cannot be

easilyunderstood,analyzed, andcaptured. Therefore, acquiringand copy-

ing these processes will be difficult for competitors. On the other hand,

these processes are a sustainableresource for competitive advantage.

This paper claims that in various organizations, there is a high vol-

ume of processes, along with a large number of process features.

Some of these processes include competitive and knowledge‐

intensive characteristics. In this situation, implementing the process

improvement plan will be a hard and complex problem. Therefore,

the main question is that how exploratory approaches (such as data

mining) can support the improvement of competitive and

knowledge‐intensive processes in a high volume of process dataset.

Overall, the main purpose of this paper is to propose a model for using

data mining to recommend the process improvement suggestions for the

competitive and knowledge‐intensive processes. For this purpose, a high

volume of process dataset is employed to implement the proposed model.

Two data mining techniques including clustering and classification are

applied to extract the behavioral patterns hidden in the large number of

processes. These patterns are employed to recommend the improvement

suggestions for the competitive and knowledge‐intensive processes.

In the following, after describing the background of this study in

Section 2, the studies related to the content of this work is reviewed

in Section 3. Section 4 explains the research methodology and

expands the proposed model. Section 5 exposes a case study of the

proposed model. Finally, Section 6 demonstrates the discussion, impli-

cation, and conclusion of the paper.

2|BACKGROUND

This section explains five main concepts employed in the proposed

model of the paper: process improvement, data mining, data mining

application for process improvement, knowledge‐intensive processes,

and resource‐based view to processes (introduced by Barney, 1991,

which emphasizes on the competitive behavior of processes).

2.1 |Processes improvement

Organizations need to improve their processes in order to successfully

achieve their intendedobjectives. In this regard, processimprovement is

one of the mostimportant dutiesof organizations in a competitive envi-

ronment. Borrego and Barba (2014) expressed that process improvement

consistsof a number activities,including designing,approving,controlling,

and analyzing the processes. For this purpose, process improvement

considers the relationships between processes, and with the human

resources, organizational context, and information resources. Damij and

Damij(2014) declaredthat processimprovement canenhance the perfor-

mance of processes from thequality, time, and cost perspectives.

2.2 |Data mining

Data mining is the process of discovering valuable patterns from the

large number of data (Koh & Low, 2004). Data mining can find and

clarify these patterns by a series of techniques in a quick way. These

124 KHANBABAEI ET AL.