Operations strategy and flexibility: modeling with Bayesian classifiers
| Date | 01 April 2006 |
| DOI | https://doi.org/10.1108/02635570610661570 |
| Published date | 01 April 2006 |
| Pages | 460-484 |
| Author | María M. Abad‐Grau,Daniel Arias‐Aranda |
| Subject Matter | Economics,Information & knowledge management,Management science & operations |
Operations strategy and
flexibility: modeling with
Bayesian classifiers
Marı
´a M. Abad-Grau
Department of Software Engineering, University of Granada,
Granada, Spain, and
Daniel Arias-Aranda
Department of Business Management, Faculty of Business Studies,
University of Granada, Granada, Spain
Abstract
Purpose – Information analysis tools enhance the possibilities of firm competition in terms of
knowledge management. However, the generalization of decision support systems (DSS) is still far
away from everyday use by managers and academicians. This paper aims to present a framework of
analysis based on Bayesian networks (BN) whose accuracy is measured in order to assess scientific
evidence.
Design/methodology/approach – Different learning algorithmsbased on BN are applied to extract
relevant information about the relationship between operations strategy and flexibility in a sample of
engineering consulting firms. Feature selection algorithms automatically are able to improve the
accuracy of these classifiers.
Findings – Results show that the behaviors of the firms can be reduced to different rulesthat help in
the decision-making process about investments in technology and production resources.
Originality/value – Contrasting with methods from the classic statistics, Bayesian classifiers are
able to model a variety of relationships between the variables affecting the dependent variable.
Contrasting with other methods from the artificial intelligence field, such as neural networks or
support vector machines, Bayesian classifiers are white-box models that can directly be interpreted.
Together with feature selection techniques from the machine learning field, they are able to
automatically learn a model that accurately fits the data.
Keywords Service operations,Bayesian statistical decisiontheory, Knowledge management
Paper type Research paper
Introduction
The useof information analysistools for supportingbusiness decisions is becominga need
for managersin the current complex andturbulent business environment(Barrientos and
Vargas, 1998). The fast progression of technical advances drives industries towards
competitionon information in order to prevent and anticipate changes in customerneeds,
technology,new industrial trends and othercompetition parameters(Anderson and Lenz,
2001). However, improvements and generalization of use of decision support systems
(DSS) have not achieved much progress. Actually, the evolution of business computing
networking and client-server architectures are impelling the utilization of shared
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0263-5577.htm
This work has been supported by the Research Program of the Spanish Ministry of Education
with the projects TIN2004-06204-C03-02 and SEJ2004-01709.
IMDS
106,4
460
Industrial Management & Data
Systems
Vol. 106 No. 4, 2006
pp. 460-484
qEmerald Group Publishing Limited
0263-5577
DOI 10.1108/02635570610661570
information in a decision support context (Burstein and the International Society for
Decision SupportSystems, 2001; Sounderpandianet al., 2005), but still m uch is to be done
to enhance learning and knowledge improvement through management information
systems. Moreover, the use of methods of formal reasoning has been scarcely applied to
scientificevidence assessing, especiallyin the field of operations management(OM) (Lotfi
and Pegels, 1996; Garbolino and Taroni, 2002). In fact, even though management
information systems literature has broadly dealt with tools to assist in managerial
decisions, thewide utility these systems generate foracademicians research seems not to
have still been discovered(Barthelemy et al., 2002).
Though, methods dealing with formal analysis are widely used to assess evidence
in empirical studies. The statistical techniques of multivariate analysis are present in
most OM studies research. However, the amount of dependencies that may exist
among different aspects of the evidence is increasing the opportunities of improving
accuracy in the results. Therefore, methods of formal reasoning have been proposed to
assist researchers in analyzing all dependencies and relationships among variables
(Garbolino and Taroni, 2002). In this context, Bayesian networks (BN) represent a
general pattern of inference, which can be suited to particular studies. By
disaggregating an inference problem into smaller “problem modules” which are
solved separately, it is possible to obtain solutions for the larger problem (Pearl, 1988).
Hence, problems occurring at the level of scientific evidence can be better understood
and then, through the use of probability calculus and statistical techniques can be
made to cohere with the entire evidential body (Neapolitan, 1990).
OM decisions are usually categorized into strategic and tactical (Schroeder, 2000).
Tactical decisions tend to be structured under a formal framework of reasoning
(Markland et al., 1998). Hence, optimization methods cover and solve most problems
with high rates of accuracy. However, strategic decisions nature is unstructured and
there is no a formal process of reasoning that suits all possible choices especially in
situations of missing data (Brown and Kros, 2003). Moreover, dependency and influence
among variables is not as clear as for tactical decisions. It is in this point where DSS play
a fundamental role in order to increase knowledge of the interdependencies among
strategic variables. Basically, a DSS can be defined as a computer system that deals with
a problem where at least some stage is semi-structured or un-structured. Usually, and
due to the ill-defined nature of information, DSS require relational database systems or
the more recent data warehouses and flexible query languages. Knowledge management
through model design demands the use of techniques from artificial intelligence and
expert systems to provide smarter support not only for decision making but also to
improve knowledge quality (Klein and Methlie, 1995). Thus, one of the main challenges
is the embracement of a much more comprehensive view of OM strategic
decisions-making DSS capable of handling less structured information and broader
concerns than mathematical models (Mintzberg and Raisinghani, 1976). In the past,
most of the research efforts in operations research focused on the development of new
algorithms to solve problems faster. In fact, important advances have been achieved in
different OM-related fields such as forecasting – see for instance Nikolopoulos and
Assimakopoulos (2003). Now, DSS applications like BN can broad knowledge in
OM decisions through new enhancement the researchers can benefit from. This new
range of possibilities motivates research on applying new DSS-based methodologies
to OM studies. The fast development of algorithms and increase on accuracy of BN
Operations
strategy and
flexibility
461
To continue reading
Request your trial