Ebook: Fusing Decision Support Systems into the Fabric of the Context

This volume contains a selection of the papers presented at the 16th IFIP Working Group 8.3 International Conference on Decision Support Systems (DSS).

The IFIP Working Group 8.3 conferences have been held biannually since 1982. The aim of these forums is to create an environment for the audience interested in current advancements in the area of Decision Support Systems (DSS) and Decision Making (DM) to come together to present their ideas on new approaches, review the latest innovations, and discuss relevant applications. The topics for each event have been specific to timely cutting-edge issues, and have included theories, systems, methodologies, algorithms, techniques, applications and technologies suitable for supporting decision making. Each IFIP WG 8.3 conference has promoted a new research theme encouraging researchers to widen the boundaries for DSS research and practice into new directions.

At the 2002 conference, held in Cork (Ireland), participants were asked to evaluate the impact of the Internet and to envisage its future potential. In Prato (Italy) in 2004, the “spirit of the humanist scholars of the Renaissance” was proposed as a source of inspiration to guide a reflection on the relevance of decision support in an uncertain and complex world. Creativity and innovation in decision making and support was the theme for the London (UK) conference in 2006. In Toulouse (France) in 2008, participants were asked to reflect on collaborative decision making, presenting latest advances and discussing the multiple facets and challenges of collaborative decision support. At the close of the first decade of the 21st century in 2010, in Lisbon (Portugal), participants were challenged to focus on the reduction of the great divide between the social activities that researchers and practitioners aim to support and those that are actually supported by DSS.

Following these traditions, it is now timely to explore the socio-technical approaches for fusing decision support, as well as opportunities and obstacles for achieving a total fusion of DSS systems in various contexts.

The conference took its prompt from the keynote by Omar El Sawy and ideas presented in his article titled “The 3 Faces of IS Identity: Connection, Immersion, and Infusion”, which originally appeared in 2003 in Communications of the AIS, volume 12. In his keynote for DSS'2010 conference, Omar El Sawy put forward the argument that “DSS has disappeared in the fabric of organizations: long live DSS”. According to this argument, DSS researchers and developers have to achieve a total alignment with the implementation contexts, both organizational and personal, in order to attain the best fit with users' preferences, behaviors and working contexts. The advantage of such a fusion view of information systems, and specifically DSS, is an opportunity for seamless integration between enterprise environments and decision support system components.

The conference solicited papers describing the application of fusion principles to DSS research and practice. The original research papers submitted to the conference covered a wide variety of topics. These topics included inter- and intra-organizational issues in fusion of DSS; fusion models for business intelligence and data warehousing; decision support fusion for organisational production planning and supply chain management; DSS fusion for emergency management; application of knowledge management for better decision support; context models for real-world decision support; intelligent approaches for fusion of decision support; fusing principles for collaborative, negotiation support and group support systems; context-based decision models, social technologies and web-based approaches in fusing decision support; incorporating complex factors in context-based decision support; fusion of decision support in specific contexts and application domains; and research methods and issues within modern DSS.

This volume of proceedings contains only research papers that were selected as a result of a review process involving at least two reviewers appointed by the program chairs. We would like to thank the reviewers for their contribution to the quality of the material presented in this book. We would also like to thank the authors for their enthusiastic contribution and effort in revising the papers to address the comments and recommendations of the reviewers.

This volume presents the most relevant and insightful research papers selected amongst the contributions accepted for presentation and discussion at the conference.

In addition to contributed papers, the program also included the following keynote presentations from the local and international distinguished researchers:

• Personalization in web search and data management, presented by Professor Timos Sellis (joint work with T. Dalamagas, G. Giannopoulos and A. Arvanitis) Research Center “Athena” and National Technical University of Athens, Greece;

• What's communication got to do with IT?, presented by Professor Dov Te'eni, Tel-Aviv University, Israel, President of the Association for Information Systems;

• Multicriteria decision support in financial decision making: An overview and a case study on bank rating, presented by Professor Constantin Zopounidis and Michael Doumpos, Technical University of Crete, Greece.

As editors, we would like to thank everyone who contributed to the content and production of this book, namely, all the authors, keynote speakers, members of the steering committee, members of the program committee, reviewers, and, last but not least, we would like to acknowledge the effort of the organizers of the conference and their university, without whom this conference would have not been possible. We would also like to thank Maarten Fröhlich from IOS Press for his availability and collaboration.

Ana Respício and Frada Burstein, DSS'2012 Program Co-chairs

Anávissos, Greece, June 28–30, 2012

This paper introduces a flexible methodology and toolset for efficient online management of intellectual capital (IC) in groups or clusters of SMEs called CROFT: CADIC Relational Online Framework and Toolset. Driven from the “bottom-up” by the requirements of the cluster participants themselves, CROFT provides a flexible framework to assist in managing the flow or transfer of IC. Based on observations of real-world SME clustering activities, CROFT has been developed to provide Group Decision and Communication Support (GDACS) to assist clusters in collective management of the group's intellectual capital through hierarchical gated access of online resources and networking coupled with a cluster-relevant IC management toolset. The pilot implementation of the system is ongoing, and while there remain issues around technical platform integration and early adoption, CROFT is designed to be flexible enough to benefit a broad spectrum of SME clusters with a wide variety of needs.

Evaluating DSS involves examining multi-dimensional attributes associated with the value of DSS though the analysis of costs and benefits, changes in the decision process, and the quality of decision support systems. In the early 1990s, DeLone and McLean presented the IS success model as a framework and model for conceptualizing and operationalizing IS (including DSS) success. The DM model is one of the widely recognized IS models based on a systematic review of 180 studies which investigated over 100 success measures. An important assumption of the DM model is that it is the model to explain IS success in a voluntary IS use context. Due to the voluntariness of IS use, the “use” of the system is in the center of the model. This paper empirically tests the validity of the DM model in an e-learning environment and found out that the DM model failed to support the positive relationships between system use and system outcome. Can the DM model of IS success be applied to evaluate the DSS success? The answer may be probably no. The future research is needed to construct a comprehensive model for evaluating DSS success. The “use” may no longer be a critical construct in evaluating DSS success in that a substantial portion of DSS published seem to be mandatory or semi-mandatory.

Medication errors are causing harm, and even death, to hospital inpatients. These preventable errors occur at the hands of the same individuals who are charged to protect and provide care to patients – healthcare professionals. While decision support technologies are available to assist healthcare providers, patients continue to experience incorrect medications, inaccurate doses/rates of medication, duplicate doses, medication interactions, and other medication errors. This paper proposes that a fusion view can provide guidance to increase the use of decision support technologies by healthcare professionals within the hospital environment to reduce medication errors and improve patient care. A case study of a hospital in the United States is used to illustrate reasons for the lack of use of existing systems and the potential of fused decision support technologies.

Organizations in various business areas operate in an increasingly regulated environment. Policy makers and regulators rely on certain assumptions rooted in various theoretical models when designing schemes to influence the decision making of economic actors. Using examples from different regulated domains this paper argues that various dimensions of a decision making situation – such as the definition of the problem, the role of participants, the strategy of the decision maker, the management of information, or the process – may be utilized by regulations or agencies as mechanisms to manipulate organizational decision making. Beyond policy design, it is contended that these dimensions can be used (1) to understand the role decision support systems play in regulated environments and (2) to make recommendations pertaining to the design and implementation of DSS dedicated to supporting managers in heavily regulated environments.

Context modeling and management are intrinsic parts of the decision-making process, especially in domains dealing with knowledge and reasoning. Actors make a decision jointly with the modeling of the context at hand. However, such a co-building of the decision-making and of the specific model of context is generally left implicit, and one generally retains {decision-making, solution} instead of {decision-making, context, solution}. Then, any reuse of a successful decision-making requires a complex phase of contextualization, decontextualization and recontextualization. This problem is well known in the scientific-workflow community—a scientific workflow being a kind of decision-mak implementation—for which we propose a solution. In this paper, we present a solution for decision-making, pointing out similarities of the approach with scientific workflows, their consideration related to task realization, and their integration in the notion of activity management. The key point is to consider decision-making through its entire processing and not its solution only. This shift of paradigm is possible, thanks to the Contextual-Graphs formalism and its uniform representation of knowledge, reasoning and context. A running example illustrates the contribution of this context-oriented approach for representing the contextualization process of decision-making.

The performance of data migration processes is an important issue when transferring data from existing source information systems to new target systems. Such a process is called ETL (Extraction, Transformation and Loading). Addressing this issue is one of the main tasks and challenges of database administrator (DBA). This activity is important because it must be accomplished as rapidly as possible to make the target system available to other systems on time. Under such a pressure, a DBA has to collaborate with several experts working on the different parts in the migration environment to solve the problem at hand. Exchange of information and knowledge between the DBA and project members becomes effective if there is a common interpretative focus and a shared context where all actors can understand each other and exchange their experiences. In this paper, we present how context sharing can help in individual decision making when dealing with ETL-processes failures, thanks to a platform called Contextual Graphs (CxGs-Platform) that allows a uniform representation of knowledge, reasoning and contexts. Our work provides a basis for the development of an experience base that will be used by a decision support system for DBA experts.

Data-mining is aimed at discovering knowledge from data. Advanced data-mining software, such as Weka, Orange and Rapid Miner, provide hundreds of data-mining methods. In order to perform a given data-mining task, these methods have to be selected and combined into a data-mining workflow. Traditionally, workflows are designed by data-mining experts, but this is difficult and there is a strong need to automate workflow design. In doing so, it is essential to be able to assess the quality of workflows. So far, this was usually assessed only through performance indicators, such as classification accuracy. In this paper, we present a workflow assessment model that uses an extended set of user-oriented indicators, which include understandability for the user, generality of used components, and robustness of the workflow. The model, which was developed using software DEXi, is qualitative, multi-attribute, hierarchical, and rule-based. We describe its components, current implementation of the model, and illustrate its performance on the case of two workflows.

Business Intelligence (BI) vendors have often asserted that the use of their tools can lead to organizational transformation. This paper compares the vendor literature on the topic with two management theories. The results of a content analysis of the vendor literature are presented, followed by an overview of dynamic capability theory and absorptive capacity. This comparison shows that the BI vendor literature treats transformation in a simplistic and narrow way. The paper argues for greater engagement between academia, BI vendors and BI customers.

This paper presents the top 20 challenges currently faced by Business Intelligence practitioners in Australia. The results were obtained from a series of nominal group technique based meetings with BI practitioners designed to elicit the most important of the challenges that they face. The most important issue that practitioners believe they face at the current time is widely seen to be developing BI systems for mobile devices.

Embedding business intelligence systems within organisations requires a seamless integration of technology, business processes and routines into the fabric of the organisation. In this paper, we propose a set of five dimensions for embeddedness of business intelligence systems within organisations. We argue that these dimensions can be related to different types of problem. We then present four case studies and a number of key insights emerge from a cross case analysis.

In the debate on the core of Information System (IS), El Sawy identified three faces of IS views: connection, immersion, and fusion. However, no research has further elaborated these concepts. For this task, we adopted the characteristics of the old concept of “emergence” to better describe and explain the shifts in IS from connection to immersion and finally towards fusion. A police organization is examined as a case study, and its Business Intelligence (BI) system and users and their interaction in solving daily tasks are observed. The study took place during two different time periods in 2009 and 2011. Both quantitative and qualitative methods were used, including a survey and interviews. Two main user groups were identified in the police organization: viewers and analysts. The results of this study showed that in the context of viewers BI is in connection view and is slowly emerging into immersion. On the other hand, in the context of analysts BI is positioned in immersion view and, based on emergent properties, we concluded that there is potential for a slow emergence towards fusion once more capabilities are integrated in the BI system used by the analysts.

This paper focuses on the adoption of the Balanced Scorecard model [1] for strategic management by a Portuguese insurance company and presents the Decision Support System (DSS) that has been conceived and developed to embody this adoption. A special emphasis is given to the definition of a strategy, the selection of the strategic measures to analyse and monitor and to the selection of initiatives taken to operationalize the identified strategy. The DSS was designed in close interaction with senior executives and was conceived aiming at enabling the company to more effectively analyse and control indicators relevant for its strategic decisions. The achievements resulting from the implementation of the system are presented and reveal that the BSC integrated with a DSS can be an effective strategic management tool for an SME.

During the last decades, the Data Warehouse has been one of the main components of a Decision Support System (DSS) inside a company. Given the great diffusion of Data Warehouses nowadays, managers have realized that there is a great potential in combining information coming from multiple information sources, like heterogeneous Data Warehouses from companies operating in the same sector. Existing solutions rely mostly on the Extract-Transform-Load (ETL) approach, a costly and complex process. The process of Data Warehous integration can be greatly simplified by developing a method that is able to semi-automatically discover semantic relationships among attributes of two or more different, heterogeneous Data Warehouse schemas. In this paper, we propose a method for the semi-automatic discovery of mappings between dimension hierarchies of heterogeneous Data Warehouses. Our approach exploits techniques from the Data Integration research area by combining topological properties of dimensions and semantic techniques.

Business analytics systems create value and provide competitive advantage for organisations. We argue that operational and dynamic business analytics capabilities can explain how value is created, sustained and renewed. We develop an evolutionary, process-oriented theoretical framework that describes how dynamic and operational business analytics capabilities interact over time to create value. We use the framework to explain how business analytics systems were integrated with the enterprise environment in a longitudinal case study of a large financial institution. A number of key factors for success with business analytics systems are identified.

In this paper, we address the problem of full option ranking in hierarchical qualitative evaluation models. Qualitative evaluation models perform partial ranking of options. The main challenge for full ranking of options with hierarchical models is finding a suitable aggregation function. A common way to define an aggregation is to use some kind of regression. Current approaches that use linear regression often fail to provide a full ranking of options in non-linear cases. Therefore we propose different methods for linear and copula-based regression functions. We evaluate two real data-mining workflows and two of its modifications, and show that full ranking of options is achieved when employing copula-based regression using Frank copula.

The Analytical Network Process (ANP) is a multicriteria decision analysis method that has been widely used the last few years to support decision making. Primary aim of this paper is to expose in detail the mathematics of the method in order to explain the mechanics of its steps and provide an exact algorithm that will give the opportunity to easily understand and implement either the method itself or modified versions of it. The paper holds an holistic approach to the method, covering not only simple ANP models but also Hierarchies and BOCR models. The practical implication of this paper is that the algorithm provided may be used as a basis for the development of custom software programs and tools that will foster the further expanding of the method.

In this paper, we present the new developments of the interactive Decision Support System (DSS) named SABILOC ([8], [1]). SABILOC is aimed at supporting decision-making concerning bicriteria location models in which the facilities to be located could have environmental impacts. The decision support is provided through two interactive phases. First, a combinatorial optimization procedure to obtain non-dominated solutions for bicriteria location problems is used. A Geographic Information System was embedded into the DSS ([7]) in order to obtain relevant data for the models concerned, especially those considering enviromental issues. In the second phase, a set of compromise alternatives can be analyzed in detail using a multi-attribute a posteriori analysis tool. This one stands for a simple interactive implementation of the conjunctive method making use of a radar chart as basis for the procedure.

We propose a Relaxed Checkpoint algorithm (R-Checkpoint) to solve Multi-Event Expert Query Parametric Estimation (ME-EQPE) problems over multivariate time series. Our proposed algorithm combines the strengths of both domain-knowledge-based and formal-learning-based approaches to learn decision parameters for yielding a reasonable time utility over multivariate time series. More specifically, our approach solves the decision optimization problems to yield the time utility from multiple decision time points, as well as learns the multiple sets of decision parameters in their respective events during the computations at a lower cost. We show that our approach produces a reasonable forecasting result by using the learned multiple sets of decision parameters.

This paper presents an innovative approach to collaborative decision making that takes into account requirements imposed by the “data deluge” era. The proposed approach builds on the synergy of human and machine intelligence. It integrates a set of interoperable services that reduce the data-intensiveness and complexity overload at critical decision points to a manageable level. Its applicability in three real-world use cases is sketched.