In today’s economy the importance of effective and efficient decision making has become
increasingly important in order to stay competitive in a global market set.
Obtaining most relevant data and outputs is the key for best decisions on every management
level. That is why the application of decision support systems (DSS) is now irreplaceable in
organizations that operate a state-of-the art decision making processes.
This paper discusses Group Support Systems, Expert Systems, Knowledge Management
Systems, Neural Networks, and Fuzzy Logic Systems focusing on
· the description of the methodology
· their applications and software vendors
· current research
· and new developments.
The analysis of these five systems is followed up by a case study that provides an example for
a real life application of a neural network. It will be exposed how this decision support system
facilitates the classification of potential patients for hospitals and how effective marketing
strategies specified to each group of patients could be created.
By comparison the application of an expert system for the same problem will be investigated.
Concluding, a summary will compose the results of this paper focusing particularly on the
outcomes of the case study. Advantages/disadvantages for the application of either system
will be examined. The authors will then eventually state their personal recommendation for
solving the problem given in the case study and comment on possible reasons why hospitals
decided to apply a neural network.
Table of Contents
1 Abstract
2 Technologies used by decision-makers in Decision Support Systems and Intelligent Systems
2.1 Group Support Systems (GSS)
2.1.1 Description of GSS Methodology
2.1.2 Applications and Software Vendors
2.1.3 Current Research Areas and New Developments
2.2 Expert Systems (ES)
2.2.1 Description of ES Methodology
2.2.2 Applications and Software Vendors
2.2.3 Current Research Areas and New Developments
2.3 Knowledge Management System (KMS)
2.3.1 Description of KMS Methodology
2.3.2 Applications and Software Vendors
2.3.3 Current Research Areas and New Developments
2.4 Artificial Neural Networks (ANN)
2.4.1 Description of ANN Methodology
2.4.2 Applications and Software Vendors
2.4.3 Current Research Areas and New Developments
2.5 Fuzzy Logic (FL)
2.5.1 Description of FL Methodology
2.5.2 Applications and Software Vendors
2.5.3 Current Research Areas and New Developments
3 Implementation Case Study for ANN Technology: Taiwanese Hospitals
3.1 Case Study: Consumers’ Behaviour in Choosing Hospitals
3.2 Problem Identification & Description
3.3 Solution of the Problem in the Case Study: ANN
3.4 Alternative Solution: Using an ES
3.5 Comparison and Recommendations
3.5.1 Critical Comparison of ANN and ES solution
3.5.2 Table Summarising Findings
3.5.3 Justification of the Organisation’s Decision
3.5.4 Recommendation
4 A Short Conclusion
References
Appendix
Appendix I: Table of Place-Time frame for Group Support Systems
Appendix II: A Processing Element
Appendix III: Different ANN topologies
Appendix IV: Results of Hospital Customer Cluster Analysis
Declaration of Academic Integrity
1 Abstract
In today’s economy the importance of effective and efficient decision making has become increasingly important in order to stay competitive in a global market set. Obtaining most relevant data and outputs is the key for best decisions on every management level. That is why the application of decision support systems (DSS) is now irreplaceable in organizations that operate a state-of-the art decision making processes.
This paper discusses Group Support Systems, Expert Systems, Knowledge Management Systems, Neural Networks, and Fuzzy Logic Systems focusing on
- the description of the methodology
- their applications and software vendors
- current research
- and new developments.
The analysis of these five systems is followed up by a case study that provides an example for a real life application of a neural network. It will be exposed how this decision support system facilitates the classification of potential patients for hospitals and how effective marketing strategies specified to each group of patients could be created.
By comparison the application of an expert system for the same problem will be investigated.
Concluding, a summary will compose the results of this paper focusing particularly on the outcomes of the case study. Advantages/disadvantages for the application of either system will be examined. The authors will then eventually state their personal recommendation for solving the problem given in the case study and comment on possible reasons why hospitals decided to apply a neural network.
2 Technologies used by decision-makers in Decision Support Systems and Intelligent Systems
2.1 Group Support Systems (GSS)
2.1.1 Description of GSS Methodology
Many decisions within organizations are made by groups in groupwork. DeSanctis & Gallupe(1987) have defined this as a decision-making group which is formed by two or more people who are responsible for detecting a problem, generating and evaluating potential solutions, or formulating implementation plan (DeSanctis & Galluple, 1987, pp. 598-609). Group decisions are seen valued more than decisions taken by single persons(Turban, Aronson, Liang, & Shard, 2007, p. 433). However, this potential is often not used due to limitations inherent in the structure and process of group decision-making (Reismann, Johnson, & Mayes, 1992, p. 169). Group Support Systems were developed to reduce and even eliminate these dysfunctions and increase benefits from groupwork.
Group Support System (GSS) is a term coined in the 1990s and was to replace a terminology called Group Decision Support System (GDSS). GSS is a branch of information technology, which combines communication, collaboration, computing and decision support technologies to facilitate the formulation and solution of unstructured problems by a group of people (DeSanctis & Galluple, 1987, pp. 598-600). It can be any kind of hard- or software that enhances groupwork. Its goal is to technically support the work of groups throughout every possible activity which was usually conducted in face-to-face activities (Turban, Aronson, Liang, & Shard, 2007, p. 453). Participants use a common computer or network to enable collaboration.
Since inputs can be anonymous, it helps workgroups quickly generate lots of ideas to solve problems or find opportunities, distil those ideas to the very best, clarify exactly what is meant, organize the ideas, evaluate and prioritize them, build consensus among the team, and finally, produce deliverables that help the team take action(GroupSystems, 2008a). Most times GSS provides fairly general support for simple activities such as idea generation, conflict resolution and voting. Though, an electronic meeting system (EMS) as a form of GSS that supports anytime/anyplace meetings provides even more additional group activities including planning of meetings, problem solving, issue discussion, negotiation, and collaboration group activities such as document preparation and sharing. Nowadays, these terms are considered to be synonymous (Turban, Aronson, Liang, & Shard, 2007, p. 453).
The University of Arizona developers claim a number of benefits for their electronic meeting system. It enables simultaneous work by all participants (e.g. human parallel processing) and provides an equal opportunity for participation. EMS encourage behaviours of participants that influence meeting productivity positively while even larger group meetings can effectively bring more information, knowledge, and skills to bear on the task. Due to the great variety of participants and improved engagement in meetings, the group can choose from a spectrum of structured or unstructured techniques to succeed in a task. EMS has access to external information which can provide the group with even more data.
A GSS supports four different kinds of possible group compositions with different technologies ranging same place/same time face-to-face meetings to virtual meetings of different places/different times. Appendix I shows the time-place matrix. Each constellation uses different techniques to improve groupwork.
2.1.2 Applications and Software Vendors
In this chapter the menu of applications and software vendors are sampled. GroupSystems (www.groupsystems.com) is the world’s largest provider of group systems to support groupwork within a focus on business environment. ThinkTank, one of GroupSystems’ products, is an interactive, web-based team-space for brainstorming, decision-making, and taking decisive action. It shortens cycle time for strategic planning, product development, problem solving, requirements gathering and other collaborative business processes. GroupSystems has also developed a software called Quickvote. This is a simple voting tool for anyone and mostly used for simple yes-or-no questions and for minor and short problems. For example, it can range from making a decision on where to go to lunch to an actual vote for politicians, heads of departments, group leaders, or other things. Its application is not restricted to any industry. For optimal usage and effectiveness, it requires quick and easy participation (GroupSystems, 2008b).
Online Workspace is an extension of Screen Sharing with the ability to share documents, files and work, etc in the same online space, but not necessarily at the same time. Software from Vignette Corp. called “Intraspect” allows users to set up workspaces. Other vendors like Microsoft (“SharePoint”), Groove Networks sell online workspace to create Web sites and post documents for people who are frequently outside a company. CollabNet, Inc. has a special workspace for collaboration of software developers.
2.1.3 Current Research Areas and New Developments
Latest developments and researches have worked on voting systems which support easy participation on mostly public and corporate elections. First trials have been conducted in the United States using electronic voting polls. The rate of invalid submissions was lower than voting paper based. Electronic mistake recognition can be implemented to decrease the rate of invalid submissions to zero. More advanced voting methods would be to access a voting tool online (Turban, Aronson, & Liang, 2005, p. 371). In addition, electronic voting saves costs of mailing, counting time and personal and printing.
Another research area involves expert/knowledge-based systems. As expert systems become more pervasive, one or more of them will be brought into meetings and discussions to assist in deliberation. They role can range from retrieving information of new alternatives to sheer helping resolve conflicts of opinion.
A third research area is composed by the adoption of Voice-Over-IP in virtual meetings. Having conferences through the Internet and adapting this to internet webcams tries to imitate face-to-face meeting. The simplicity and easiness in not having to leave one’s desk to speak to others over a long distance without using a costly phone is the optimal solution of Voice- Over-IP. Additionally, voice messages left on a team-space on various topics enables other participants to simply listen to statements and contributions easier. Recording a voice message is also much quicker then writing a message. Companies and internet site that offer internet telephony are www.pc-telephony.com and Siemens Communication on www- communications.usa.siemens.com
Virtual organizations have aroused and constitute a current and ongoing field high use of group support systems. The organization holds a common identity. The virtual organization is often described as one that is completely designed to solely have geographically spread, functionally and/or culturally diverse entities of which all are connected via electronic forms of communication. Managerial executives, employees, and customer are all physically separated. The ‘company without walls’(Galbraith 1995) basis its functionality on collaborative networks of people working together, regardless of location.
2.2 Expert Systems (ES)
2.2.1 Description of ES Methodology
An expert system, also called knowledge based system, is a computer program that simulates the judgement and behaviour of a human or an organization that has expert knowledge and experience in a particular field(SearchCIOMarket.com, 2007). with the help of an expert system it is possible to solve problems which normally requires human expertise. Expert systems allow ‘attaining high-level decision performance in a narrow problem domain’(Turban, Aronson, & Liang, 2005, p. 549).
There are four features for expert systems: Expertise, symbolic reasoning, deep knowledge and self-knowledge. The expertise of experts is needed as an input for expert systems, so that the software can make expert-level decisions. The second feature, symbolic reasoning is rather used as mathematical calculations. For the development of an expert system there is a great need of deep, complex knowledge, the knowledge of a non-expert is not deep enough. The last feature, self-knowledge means that an expert must be able to explain somebody how a conclusion was reached. It is also important that an expert learns from his/her success or failure(Turban, Aronson, & Liang, 2005). Satz überarbeiten
Expert systems belong to the knowledge stage of the evolution of artificial intelligence. In this stage artificial intelligence gets integrated in real-world applications, normally in a ‘narrowly defined domain with specialized knowledge (Turban, Aronson, & Liang, 2005, p. 542). The knowledge base, the nerve system of an expert system is besides the inference engine and the user interface, the main component of expert systems. The inference engine is the computer program of the expert system that provides a methodology for reasoning about information in the knowledge base as well as on the blackboard and for formulating conclusions. The inference engine shows the user how to use the saved knowledge; hence it is also called the brain of the expert system(Turban, Aronson, & Liang, 2005). The third main component, the user interface is normally a question-and-answer approach, which allows friendly, problem- oriented communication between the user and the computer. The user interface is supported by a language processor. when considering an expert system, there are two different environments: the development environment and the consultation environment.In the development environment the knowledge of an expert, also called expertise, is put into the knowledge base after the Expert Systems builder has built its components. In the second environment, the consultation environment, a non-expert uses the expert system for making decisions which are based on the knowledge of an expert . This knowlege was saved in the development environment beforehand (compare figure 10.3., p.555, Turban, Aronson, & Liang, 2005).
Knowing the main facts about the development environment and the consultation environment is important for understanding how an expert system works. In the knowledge base the knowledge is represented in a computer-understandable format and than used by the inference engine to infer conclusions from existing facts and rules. The expert knowledge can be presented in different ways, for example in rule-based systems, implying that there is a rule which says what happens, in the case of something (if ... then ...). With regard to complex decisions it isn't possible to use only one rule. If you chain multiple rules beaded on available data it is called inference, this process is done by the inference engine. Incensing is possible as forward chaining and as backward chaining. Looking at the if part of a rule first it is called forward chaining looking at the then part first it is called backward chaining.
2.2.2 Applications and Software Vendors
Expert systems have been applied in many business areas in the past, for example in finance, data processing, marketing, human resources, manufacturing and medicine(Turban, Aronson, & Liang, 2005; Pfeffer, 2002). One popular example is Dendral. Dendral has used knowledge- or rule-based reasoning commands to predict molecular structure of organic chemical compounds of known chemical analyses and mass spectrometry data.
Another example is the Microsoft Windows operating system troubleshooting software, which is located in the help section of the taskbar. There you can search for different solutions if you have a problem or want to know something. inserting a keyword, will make the software search for solutions which are linked to the keyword and saved in the knowledge base. The users will than get a number of possible results and can choose which result is the right one.
In the automotive sector expert systems help mechanics in diagnosing for maintenance and repair. The technical details, experiences and important information are saved in the expert system. DAX is an example for the use of an expert system in the automotive sector; it is used by Daimler for the control of the quality of automatic gearboxes. The function of the gear boxes is checked and the expert system shows three different diagnoses:
- okay, if there are no mistakes in the gearbox
- if the proofed data differ from the approximate value
- if nothing above is the case, you get a default diagnosis but no more advices(Community of Knowledge, 2004).
Another application of an expert system is OncoLogicTM, which evaluates the likelihood that a chemical may cause cancer. The program, which runs on each Windows computer, was released by EPA at no cost and can be used by any researcher or organization. Onco LogicTM determines the cancer-causing potential by structure activity relationship1 analysis, involving the decision logic of human experts and incorporation knowledge of how chemicals cause cancer in humans and animals(Environmental Protection Agency, 2008).
2.2.3 Current Research Areas and New Developments
Current expert systems are really specialized, that means that they cannot be used for a lot of different tasks, because the knowledge base is quite small. In the future expert systems will get much more important for saving money in the long run and adjusting a great need for experts(AAAI, 2007). expert systems will get more expertise as input so that the knowledge base will grow, the system will also deal with domains of wide specialization. Data of the past must be stored and evaluated for getting enough information and having the possibility to develop a greater knowledge base(Lueg, 2001).
Regarding medicine expert systems will get much more important for diagnosis; hospitals often use expert systems for analysing data in the laboratory. Besides the time saving, the easy usage of expert systems in this domain is also an important point, because the expert systems allow faster diagnosing(Pfeffer, 2002).
2.3 Knowledge Management System (KMS)
2.3.1 Description of KMS Methodology
Knowledge management is defined as the practice of selectively applying knowledge from previous experiences of decision-making to future decision making actions with the effective purpose to enhance organisational effectiveness (Jennex, 2006). KM systems entirely combine organizational and technical solutions in order to attain the goals of knowledge storage and reutilization to ultimately enhance decision making processes in organisations. A KMS serves as an integrated service to deploy KM instruments for networks of stakeholders - predominantly the active knowledge workforce - in knowledge-significant business operations along the complete knowledge life cycle. KMS can be applied for a variety of collaborative, cooperative, adhocracy and hierarchy societies, virtual organizations, communities and other virtual networks, to manage media contents, activities, interactions and work-flows purposes, works, projects, networks, branches, privileges, roles, participants and other active individuals(Maier, 2007). The ultimate aim is to extract and produce essential knowledge and to improve, leverage and transmit new outcomes of knowledge providing innovative services using new formats and interfaces and different communication channels(Marwick, 2001).
KMS are based on the perception of a persisting knowledge cycle model in organisations (Alavi & Leidner, 2001) consisting of six stages: knowledge creation, capture, refining, knowledge storage, management, and knowledge application. KMS apply technologies to support the continuous refining of knowledge in a constantly changing organisational environment. Knowledge management systems specifically support the various stages.
Information technology components that are the essential basis for a successfully functioning KMS can be divided into the following three categories:
Firstly, communication technologies enable individuals to access relevant and specific knowledge and to communicate with corporate field experts. The internet, corporate intranet systems, e-mail clients and various web-based tools provide the required communication technology. Secondly, collaboration technologies enable the corporate community to work jointly from different places. Specified groups of individuals get the possibility to edit work documents synchronously and asynchronously in a collaborative manner. Moreover, various other collaborative computing activities like electronic brainstorming improve KM - team work. Thirdly, Storage and retrieval technologies serve as important components for storing and the retrieval of knowledge that is accumulated in the process of knowledge management. An extensive set of applications like document-management systems and specialised storage systems (data repository) that are accessible from a variety of sources play an important role in modern KMS(Turban, Aronson, & Liang, 2005). These three technological components are significant requirements for the implementation of a KMS in an organisation and a must for companies that want to profit of KM.
2.3.2 Applications and Software Vendors
Experts in the field of Decision Support Systems generally define KMS rather as a methodology used to leverage business practices than a specific technology. Therefore the following paragraphs illustrate significant technological solutions in the ambit of KM and provide valuable information about leading vendors and manufacturers of KMS solutions.
Knowledge Servers are servers containing fundamental KM software including knowledge repositories and access enablement. These servers create interconnectivity between individuals, users and content and lead to the integration of collaboration, personalization and retrieval features. The immediate access to information via a knowledge repository and intelligent retrieval mechanisms allows an efficient distribution of time-sensitive information. In the ambit of KMS Sequoia Software XML Portal Sever, Intraspect Knowledge Server and IDOL are dominant products of leading manufacturers.
Enterprise Knowledge Portals (EKP) are efficient tools to configure KMS combining data integration, collaboration and reporting mechanisms. The portal aggregates holistically information needs of the individual like web mail, dynamic feeds, calendar tasks, web links etc. The leading EKP vendors are Brio, IBM, Oracle, Dataware and Sybase. Collaborative Computing Tools (CCT)are tools with the primary task to leverage teamwork with electronic idea categorization embedded in an electronic brainstorming system. IBM’s Lotus Notes/Domino and Latitude’s Meetingplace are the leading CCT tools. Electronic Document Management (EDM)empowers an organisation to introduce the format of electronic documents that can easily be retrieved by web-based applications. As a result EDM systems lead to enhanced document management, more efficient operations and a collaborative approach towards document development. EDM systems like Xerox’s DocuShare, Estman’s Enterprise Work Management and IBM’s Lotus Notes are good examples in this ambit.
Search Engines constitute significant tools to locate, index, catalogue specific data from huge corporate repositories. Software vendors like Microsoft, IBM, Google and Inktomi offer valuable solutions that can clearly leverage KM.
Knowledge Harvesting Tools (KHT)are valuable tools that provide the advantage of capturing tacit knowledge without the necessity of involving the contributor essentially. KHT software analysis, e-mails and various other document types of field experts in order to parse specific subject know-how and to make the information available to users within the organisation. Leading vendors of KH applications are Oracle, Microsoft and KnowledgeX. Knowledge Management Suites constitute holistic knowledge management solutions for companies integrating storage technologies, collaboration and communication technologies into a single package integrated in internal and external data repositories. The leading provider of knowledge management suites is IBM including products like Websphere, Sametime and Domino. Various other Software vendors from the ERP sector like Peoplesoft and SAP also offer out-of-the-box solutions(Turban, Aronson, & Liang, 2005).
2.3.3 Current Research Areas and New Developments
An important area of research in the ambit of KMS is the development of Knowledge Management 2.0. It can be described as the expansion of existing KMS to Web 2.0 communities, blogs and websites that are commonly used by industry specialists. This web based unstructured and globally dispersed knowledge is recognised as an essential resource that future KMS applications need to exploit beyond their own company intranet to retain competitive advantages in the times of Web 2.0(Spanbauer, 2006).
Furthermore, the development of intelligent Multimedia Retrieval Systems is becoming a key issue in KM research. Modern multimedia retrieval systems enable companies, through intelligent recognition of speech and video files, to produce textual transcriptions allowing users to query this unstructured knowledge by keywords. The field of real-time analysis of spoken discourse as a means of meeting support is also becoming a focus of research in KM. Prototype programmes like IBM’s MeetingMiner capture and analyse audio files and produce textual records. The aim of researchers is to revolutionise KM by real-time support during meetings by automatically bringing up significant information related to discussions.
The integration of KMS with other enterprise-wide information systems is also a big issue. Developments that integrate KM in persisting system like Customer Relationship Management, Business Intelligence, Supply Chain Management have become essential(Turban, Aronson, & Liang, 2005).
2.4 Artificial Neural Networks (ANN)
2.4.1 Description of ANN Methodology
Artificial Neural Networks (ANNs), or neural networks, according to Turban, Aronson, &Liang (2005, p. 547) are a ‘set of mathematical models that simulate the way a human brain functions’. Despite their name, they do not copy the physical structure of a human brain. The analogy rather stands for the creative problem solving approach and the information processing system inherent to a human brain that is employed in ANNs.
[...]
1 Technique to correlate the biological activity of a chemical to its structure (e.g. shape, size, chemical arrangement, distribution of functional groups)
-
-
-
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X.