Knowledge Discovery Process and Methods to Enhance Organizational Performance

239

Chapter 12

Applications of Data

Mining in Organizational

Behavior

Arash Shahin and Reza Salehzadeh

Introduction

Data mining has already been successfully applied in many domains. Research has

shown that data mining techniques have the most applications in the eld of market-

ing (e.g., Chang and Wand 2006; Lu et al. 2012; Sahi et al. 2012). In addition, many

studies have been conducted in accounting and nancial management (e.g., Lam

2003; Laitinen 2005; Gregoriou and Pascalau 2012); human resource management

activities (see the literature review performed by Strohmeier and Piazza 2013); opera-

tions research (Meisel and Mattfeld 2010; Corne et al. 2012); industrial manage-

ment (Braha 2001; Ni et al. 2007); and so on. However, one domain in which data

Contents

Introduction ......................................................................................................239

Association Rules Mining ..................................................................................240

Articial Neural Networks ................................................................................ 246

Clustering .........................................................................................................250

Decision Tree ....................................................................................................253

Conclusions ......................................................................................................255

References .........................................................................................................255

240 ◾ Arash Shahin and Reza Salehzadeh

mining has rarely been applied is organizational behavior. Organizational behavior

is a eld of study that investigates the impact that individuals, groups, and structure

have on behavior within an organization, for the purpose of applying such knowl-

edge toward improving an organization’s eectiveness (Robbins and Judge 2012).

Data used in data mining is usually from a previously prepared database but

data employed in the eld of organizational behavior is obtained from question-

naires. erefore, questionnaire data will have to be prepared for use in data min-

ing. Now the question is “Can we use the data obtained from questionnaire and

through self-reported for data mining too?” It seems that the application of data

mining in this eld can be eective for achieving better results. Questionnaires are

usually analyzed statistically using hypothesis testing, which are either conrmed

or rejected. Data mining techniques have the ability to seek knowledge (which is

previously unknown) from data, and thus gives the analyst insights for future deci-

sion making. On the other hand, organizational studies seek to explain, predict,

and control employees’ behaviors and for such purposes, data mining techniques

can be very helpful.

e objective of this chapter is to investigate the four dierent data mining

techniques utilizing data from questionnaires in the eld of organizational behav-

ior. e examples shown in each part will reveal the eective role of data mining in

organizational behavior studies.

Association Rules Mining

Association rules mining (ARM) was rst introduced by Agrawal et al. (1993) and

had received a great attention. Association analysis is a type of undirected data min-

ing that nds data patterns. Using association rules, dependencies and relationships

can be discovered among existing data in a database.

Association rules (AR) can be utilized in two ways for mining data from a question-

naire. e rst method involves investigating all possible relationships between ques-

tions under the variables of interest. For instance, consider a questionnaire for assessing

a variable of interest, “self leadership.” Using the association rules, the relationship

among self-leadership strategies can be investigated. Possible questions under this vari-

able will ask respondents about self-goal setting and self-cueing. AR will then be used

to assess relationships between the two questions (Shahin and Salehzadeh 2013).

e second method of utilizing AR in questionnaire data is to investigate rela-

tionships between demographic features and the variable of interest. For instance,

the relationship among demographics such as age, education, and gender and the

self-leadership strategies can be investigated, in order to nd if the ratio of age, edu-

cation, and gender type inuence the ratio of self-leadership strategies. For a better

understanding of the subject, consider the following example:

207 questionnaires including questions about the self-leadership strategies

(behavior-focused strategies, natural reward strategies, and constructive thought