Showing papers in "International Journal of Information Technology and Decision Making in 2002"
TL;DR: A promising approach of data mining to classify the credit cardholders' behavior through multiple criteria linear programming is provided, using the well-known commercial software package SAS to implement this technology by using a real-life credit card data warehouse.
Abstract: Data mining becomes a cutting-edge information technology tool in today's competitive business world. It helps the company discover previously unknown, valid, and actionable information from various and large databases for crucial business decisions. This paper provides a promising approach of data mining to classify the credit cardholders' behavior through multiple criteria linear programming. After reviewing the history of linear discriminant analyses, we will describe first a model for classifying two-group (e.g. bad or good) credit cardholder behaviors, and then a three-group (e.g. bad, normal, or good) credit model. Besides the discussion of the modeling structure, we will utilize the well-known commercial software package SAS to implement this technology by using a real-life credit card data warehouse. A number of potential business and financial applications will be finally summarized.
126 citations
TL;DR: Two open topics are suggested: the anti-accuracy principle and the digital image compression technique based on the fuzzy if-then rules extracted by using information matrix technique.
Abstract: Strong interests in the small-sample problem have been given towards for establishing several information diffusion techniques for pattern recognition. In this paper, we review and formalize three techniques: the soft histogram, the self-study discrete regression, and the interior-outer-set model. To promote the development of this area, in this paper we suggest two open topics: the anti-accuracy principle and the digital image compression technique based on the fuzzy if-then rules extracted by using information matrix technique.
60 citations
TL;DR: In this article, the authors look at knowledge as a manifest ability of purposeful coordination of action and redefine the purpose of knowledge management as turning information (description) into knowledge (action) and not vice versa.
Abstract: Knowledge Technology (KT) is an important new development, extending and ultimately replacing IT. Meaningful and substantial Knowledge Management (KM) is crucially dependent on a useful and operational definition of knowledge. Such notion of knowledge must be clearly differentiated from so called "explicit (or codified) knowledge", i.e. from information. Information, in any form or shape, is not knowledge. While information is a symbolic description of action, knowledge is action itself. The understanding that knowing is doing and doing is knowing comes from the Western philosophical tradition of pragmatism, exemplified by Dewey, Lewis and Polanyi. In this paper, we look at knowledge as a manifest ability of purposeful coordination of action and redefine the purpose of knowledge management as turning information (description) into knowledge (action) and not vice versa. While there can be an information overload, there is never any "knowledge overload".
42 citations
TL;DR: The interactive system PROMOIN has been designed in order to work with interactive techniques for Linear Multiobjective Programming problems, and has been implemented under Windows environment, with the aim of providing the user with a friendly interface.
Abstract: The interactive system PROMOIN is presented in this paper. This system has been designed in order to work with interactive techniques for Linear Multiobjective Programming problems. The main interactive procedures of the literature have been incorporated into the system, as well as the possibility to change between methods along the solution process, if the user wishes so. This change-of-method option has been developed on a theoretical base devoted to transfer information from one method to the other, so that all this information is not lost when changing to another algorithm. The program has been implemented under Windows environment, with the aim of providing the user with a friendly interface.
27 citations
TL;DR: By using two independent variables of preceding periodical data and index of time, the developed model not only shows the pattern of the seasonal variation, but also the yearly trend.
Abstract: Fuzzy regression model is an alternative to evaluate the relation between independent variables and dependent variable among the forecasting models when the data are not sufficient to identify the relation. Such phenomenon is significant especially for seasonal variation data for which large amount of data are required to show the pattern. However, few researches have been done on this issue. Because of its increasing importance in industries, in this study, we propose a method of applying fuzzy regression model for this purpose. By using two independent variables of preceding periodical data and index of time, the developed model not only shows the pattern of the seasonal variation, but also the yearly trend. From the results of the illustration, the average forecasting error is below 1.85% which, in comparison to the most commonly used Quadratic Trend Analysis of 2.91% and the Double Exponential Smoothing Model of 4.29%, has a better performance.
26 citations
TL;DR: A license management model to support not only typical music distribution models from music distributor to consumer (B2C), but also peer consumer to peer consumer (C2C) distribution model is proposed.
Abstract: The popularity of peer-to-peer (P2P) file sharing software has created a pressing need for digital music providers to resolve copyright and piracy problems of disseminating digital music through the Internet. Digital Rights Management (DRM) technology ensures protection of rights for distributed music over the Internet. Existing DRM systems emphasize protection based on business-to-consumer (B2C) distribution model, but do not tackle protection on a consumer-to-consumer (C2C) distribution model. This paper proposes a license management model to support not only typical music distribution models from music distributor to consumer (B2C), but also peer consumer to peer consumer (C2C) distribution model. The proposed license management model consists of two types of licenses, namely official licenses and peer licenses. These licenses are to be implemented in XrML. The concepts of these licenses facilitate both B2C and C2C distribution models. Functionality tests of the proposed model with certain common peer-to-peer sharing activities have been conducted.
25 citations
TL;DR: A multiobjective modelling approach that interactively assists IT managers in their attempts to reduce a given risk by evaluating and selecting portfolios of security measures and establishing the "best" one with respect to the decision-maker's preferences is presented.
Abstract: Security issues in information management are increasingly moving towards the centre of corporate interests. This paper presents a multiobjective modelling approach that interactively assists IT managers in their attempts to reduce a given risk by evaluating and selecting portfolios (i.e. bundles) of security measures. The proposed multi-step procedure identifies attractive portfolio candidates and finally establishes the "best" one with respect to the decision-maker's preferences. Our model and its possible application are demonstrated by means of a numerical example based on real-world data that evaluates the risk of hacking faced by a Local Area Network in an academic environment.
24 citations
TL;DR: To solve fuzzy and challenging problems, an effective concept and model of competence set analysis is introduced and it is shown how IT can help in these three types of problems.
Abstract: In this paper, we discuss how information technology (IT) affects and influences people to make decisions We first introduce human behavior mechanism and habitual domains — the software that drive the behaviors Then we discuss the impacts of IT on decision elements and environment, and then IT's impacts on a variety of decision problems including routine problems, mixed routine problems, fuzzy problems and challenging problems IT is useful in solving routine problems but not as obvious in solving fuzzy and challenging problems To solve fuzzy and challenging problems, an effective concept and model of competence set analysis is introduced Finally, we describe three types of competence set analysis and show how IT can help in these three types of problems
20 citations
TL;DR: The new method for the construction of exact copy of expert knowledge-CONSER is described and gives new information about the striking ability of expert to recognize and classify the situations on the basis of their knowledge.
Abstract: The problems of the nature of expertise and the possibility to create a copy of expert knowledge in the computer are the main themes for the paper. In the first part of this paper, a short survey of existing knowledge about expert behavior is given. The new method for the construction of exact copy of expert knowledge-CONSER is described. The utilization of new method gives new information about the striking ability of expert to recognize and classify the situations on the basis of their knowledge.
16 citations
TL;DR: The interesting finding is that parameter "k" affects the allocation efficiency and the trader ratio in continuous double auction markets.
Abstract: Replacing human traders by "zero-intelligence" (ZI) programs that submit random bids and offers in their experiments, Gode and Sunder11 conclude that the primary cause of high allocation efficiency and the convergence to theoretical equilibrium price in a continuous double auction market are the market discipline rather than the intelligence of traders. In this paper, we study the influence of ZI traders' expected profit margin on the continuous double auction markets. For this purpose, we develop a general zero-intelligence model that is called "k-ZI" where k ∈ [0, 1]. When k = 0, a k-ZI trader becomes a ZI with constraint (ZI-C) trader; when k = 1, a k-ZI trader becomes a truth-telling trader. The parameter "k" determines the expected profit margin of k-ZI traders. By the experiments conducted, we investigate its influence on the continuous double auction market, such as on the allocation efficiency, the trade ratio and the average trade price. The interesting finding is that parameter "k" affects the allocation efficiency and the trader ratio in continuous double auction markets. In order to interpret experimental results economically, the anti-Marshallian paths of double auction markets are provided.
15 citations
TL;DR: Both logistic regression and neural networks have their advantages and disadvantages and the combined of these two can enhance their predictive ability and offset their weakness and the power of the present proposal in a real data set.
Abstract: In credit risk management, the on-line analytical process has been accepted by most credit card issuers. The major tools used in such an OLAP are statistics and neural networks. Through a designed algorithm, the OLAP generates scores for each account or for each customer which depends on the level of the processing. Generally speaking, logistic regressions and feed-forward networks are the major players in OLAP of this field and usually are used separately. This paper discusses an approach — Dual-Model Scoring System — to combine these two major players and use them together in the credit scoring. Primarily, the classification problem for two classes are considered. By the Bayesian rule, the objective function of classification can be reduced to estimate the Bayesian posterior probability. Such a probability is estimated by using the MLE approach in logistic regressions and the Two-Stage (Gibbs) learning algorithm3 in feed-forward networks. The motivation of the proposal comes from the following two considerations: (1) Both logistic regression and neural networks have their advantages and disadvantages and the combined of these two can enhance their predictive ability and offset their weakness. (2) To reduce the false positive rate in the decision region. Besides the discussion of the architecture design of Dual-Model Scoring System, the paper has demonstrated the power of the present proposal in a real data set.
TL;DR: It is shown that context-specific independence can be used to further decompose some of the conditional probabilities and provided, for the first time, empirical evidence that demonstrates the computational benefits of exploiting context- specific independence.
Abstract: Three kinds of independence are of interest in the context of Bayesian networks, namely conditional independence, independence of causal influence, and context-specific independence. It is well-known that conditional independence enables one to factorize a joint probability into a list of conditional probabilities and thereby renders inference feasible. It has recently been shown that independence of causal influence leads to further factorizations of some of the conditional probabilities and consequently makes inference faster. This paper studies context-specific independence. We show that context-specific independence can be used to further decompose some of the conditional probabilities. We present an inference algorithm that takes advantage of the decompositions and provide, for the first time, empirical evidence that demonstrates the computational benefits of exploiting context-specific independence.
TL;DR: The proposed agent fabrication scheme brings forth a potential solution for creating agents in agent-based e-commerce applications and its implementation in multi-agent systems based on the SAFER architecture.
Abstract: In the last decade, agent-based e-commerce has emerged as a potential role for the next generation of e-commerce. How to create agents for e-commerce applications has become a serious consideration in this field. This paper proposes a new scheme named agent fabrication and elaborates its implementation in multi-agent systems based on the SAFER (Secure Agent Fabrication, Evolution and Roaming) architecture. First, a conceptual structure is proposed for software agents carrying out e-commerce activities. Furthermore, agent module suitcase is defined to facilitate agent fabrication. With these definitions and facilities in the SAFER architecture, the formalities of agent fabrication are elaborated. In order to enhance the security of agent-based e-commerce, an infrastructure of agent authorization and authentication is integrated in agent fabrication. Implementation and prototype applications show that the proposed agent fabrication scheme brings forth a potential solution for creating agents in agent-based e-commerce applications.
TL;DR: In this article, the authors propose a procedure that incorporates multiple criteria decision making (MCDM) into system dynamics modeling to handle dynamic multiple criteria situations, which they name dynamic MCDM.
Abstract: Many societal decisions involve complexity and conflicting objectives. Preferences in such environments can be expected to change as situations evolve. In this paper, we propose a procedure that incorporates Multiple Criteria Decision Making (MCDM) into system dynamics modeling to handle dynamic multiple criteria situations, which we name dynamic MCDM. A case of urban infrastructure is presented to illustrate the procedure. Dynamic MCDM can handle different lags in economic, social, economic and technical effects of large scale systems. Thus, it may help decision makers avoid selecting alternatives apparently effective in the short term, but detrimental in the long term.
TL;DR: A detailed model for designing an online Multi-Criteria Electoral System (MCES) is presented, based on emerging technology and the Analytic Hierarchy Process, to minimize the imperfections of traditional voting systems and bring about electoral reform.
Abstract: A detailed model for designing an online Multi-Criteria Electoral System (MCES) is presented. Such a system is based on emerging technology and the Analytic Hierarchy Process (AHP). It deals with the intensity of preferences in contrast with traditional "yes-no" votes. Our objective is to minimize the imperfections of traditional voting systems and bring about electoral reform. A prototype Web-based system is developed and then used to analyze a real public policy issue. We also explore the practical implications of such a system in the context of a few other scenarios.
TL;DR: It is concluded that politically sponsored bandwagon often leads to a quick establishment while the market-sponsored bandwagons tend to accommodate market and technology dynamics, at least for the wireless industry.
Abstract: Standardization development and the global standards adoption are in the center of competition for the global mobile communication industry This study conducts an historical analysis of three generations of standardization in mobile communications in the world It aims to examine how the corporation strategy, institutional collaboration, and government policy have influenced the outcome of standardization It focuses on how competing standards are adopted and deployed in national and global markets Empirical data has been presented to test the theory of bandwagons in standard adoption by using ratios of total global users to total domestic/home users Findings lead to the conclusion that standards with well-established domestic bases often start their global bandwagons with a bandwagon ratio between 12 and 23 and the global competition is centered between the politically sponsored European bandwagon and the market-sponsored US bandwagons It concludes that politically sponsored bandwagon often leads to a quick establishment while the market-sponsored bandwagons tend to accommodate market and technology dynamics, at least for the wireless industry
TL;DR: It is claimed that there is a certain structure present in a VO that can be found in specific roles identified in two VO case studies, which are within knowledge teams around a specific knowledge field, who cooperate with other knowledge teams.
Abstract: Virtual organizations (VO) are often characterized by their loose structure, lack of hierarchy and their empowered geographically dispersed members. However, in this paper it is claimed that there is a certain structure present in a VO. This structure can be found in specific roles identified in two VO case studies. The roles are within knowledge teams around a specific knowledge field, who cooperate with other knowledge teams. A second role can be recognized in a management part of the VO. Management is seen as an architect role, where the content of the work is integrated into an overall picture, and a broker role, who mediates between the different parties involved. The last role contains the autonomic partners who are the sponsors of the VO. Furthermore, the power of the VO is not equally distributed among its members, but divided over groups that are part of the architect management role.
TL;DR: This paper studies the class of signaling games in which a unique perfect Bayesian Nash equilibrium outcome satisfying the Intuitive Criterion exists and proves three basic theorems leading to the proof of the main theorem that gives sufficient conditions under which a signaling game has a unique outcome.
Abstract: In this paper, we study the class of signaling games in which a unique perfect Bayesian Nash equilibrium outcome satisfying the Intuitive Criterion exists. We first describe the models of signaling games and lay out the definition of equilibrium and some basic assumptions. We then prove three basic theorems leading to the proof of the main theorem that gives sufficient conditions under which a signaling game has a unique outcome. The proof of the main theorem also leads to a method for computing the equilibrium outcome. We formally present the algorithm and illustrate its application through a numerical example. Finally, we apply our algorithm to enterprise management decision-making.
TL;DR: A case study at a large tertiary care hospital discusses a number of issues that arise in analyzing occupancy data which have implications for design of healthcare operations oriented data warehouses and analysis tools.
Abstract: Managerial decision making problems in the healthcare industry often involve considerations of customer occupancy by time of day and day of week. Through a case study at a large tertiary care hospital, we discuss a number of issues that arise in analyzing occupancy data which have implications for design of healthcare operations oriented data warehouses and analysis tools. We offer practical solutions to these problems including a transaction oriented database design, a general database framework and software tool for analysis of occupancy related data and a method for simulating entity flow from the data mart.
TL;DR: The design of two kinds of handwritten Chinese character recognition systems based on totally parallel combination, in which human intelligence and computer capabilities are combined together through multi-step supervised learning.
Abstract: The metasynthetic approach for solving complicated problems was proposed in 1990.1 And the characteristics of metasynthetic approach can be summarized as human-machine cooperation and integration. Directed by the idea of metasynthesis, the design of two kinds of handwritten Chinese character recognition systems are given in this article. All the designs focus on incorporating human knowledge into multiple classifier combination, which is different from conventional integration. The first one is multi-stage adaptive weighted multiple classifier combination, in which a neural network for coefficient predicting is trained by supervised learning to provide weights suitable for the input pattern. And the second scheme is based on totally parallel combination, in which human intelligence and computer capabilities are combined together through multi-step supervised learning. The experimental results demonstrate substantial improvement in overall performance for handwritten Chinese character recognition with thousands of classes that must be discriminated.
TL;DR: This paper presents how the different types of data mining rules can be transformed into the domain knowledge of any general-purpose expert system, and presents how certain attribute values given by the user as facts and/ or goals can determine, through a forward and/or backward chaining, the related datamining rules.
Abstract: Data mining is an emerging research area that develops techniques for knowledge discovery in huge volumes of data. Usually, data mining rules can be used either to classify data into predefined classes, or to partition a set of patterns into disjoint and homogeneous clusters, or to reveal frequent dependencies among data. The discovery of data mining rules would not be very useful unless there are mechanisms to help analysts access them in a meaningful way. Actually, documenting and reporting the extracted knowledge is of considerable importance for the successful application of data mining in practice. In this paper, we propose a methodology for accessing data mining rules, which is based on using an expert system. We present how the different types of data mining rules can be transformed into the domain knowledge of any general-purpose expert system. Then, we present how certain attribute values given by the user as facts and/or goals can determine, through a forward and/or backward chaining, the related data mining rules. In this paper, we also present a case study that demonstrates the applicability of the proposed methodology.
TL;DR: The model is formally specified in terms of price, profit, and group gaining functions and is applied to some examples of societies of selfish agents in antagonistic groups to illustrate its dynamic properties.
Abstract: A model of surplus values within information ecosystems is presented. The model is based on the classical definition of surplus value. However, as this definition was developed within a manufacturing industry context, some modifications are necessary to adopt it to the context of information ecosystems, e.g. by taking into account that products are "virtual" rather than physical. Just as in agent-based computational economics, here the economics is modelled as evolving systems of autonomous interacting agents in an evolutionary framework. In this way the resulting model is able to capture more dynamic scenarios. The model is formally specified in terms of price, profit, and group gaining functions and is applied to some examples of societies of selfish agents in antagonistic groups to illustrate its dynamic properties. Moreover, the paper show how the model builds upon labour theory of value and contrasts it to consumer value models.
TL;DR: A simple reformulation that enables the binary assignment problem with a side constraint requiring the objective function to receive a specified value to be solved by commonly available heuristic methods is presented.
Abstract: The binary assignment problem with a side constraint requiring the objective function to receive a specified value, which in general is an NP-hard problem, has been the focus of several papers in recent years. The current literature addresses various theoretical aspects of the problem, with a particular emphasis on a simplifying special case, but stops short of giving any computational experience. In this paper, we present a simple reformulation that enables the problem, and some of its extensions, to be solved by commonly available heuristic methods. We present preliminary computational experience with a Tabu search method that illustrates the effectiveness and computational robustness of the approach.
TL;DR: This paper alters the formulation by adopting assumptions suitable for electronic trading environments, and then derive the optimal trading strategy in terms of the agent's expected utility, in order to reduce the infinite state space into the one with a manageable size.
Abstract: The recent advent in Internet technology has made electronic trading increasingly popular. Nevertheless, existing electronic trading systems still heavily rely on human decision making. In order to facilitate the new trading environment, one of the main trends is to employ autonomous agents as representatives of human buyers and sellers. In particular, given a user requirement and an imposed deadline, an autonomous agent has to search for possible deals from an electronic marketplace on behalf of its owner. One problem such an agent must face is: given a collection of offers and the remaining time, should I accept the current best offer, or continue to search for a better one, at a risk of losing the current offers? Similar time-constrained trading problems have been studied long time ago by the Markov decision process community. This paper alters the formulation by adopting assumptions suitable for electronic trading environments, and then derive the optimal trading strategy in terms of the agent's expected utility. Three optimization methods are proposed in order to reduce the infinite state space into the one with a manageable size. Experimental results verify the effectiveness of the methods for practical use.
TL;DR: Insight on the essence of virtual organizations is presented, which is IT enabling, and a brief overview is given upon the papers that are included in the special issue.
Abstract: Firstly, insight on the essence of virtual organizations is presented, which is IT enabling. Then, a brief overview is given upon the papers that are included in the special issue.
TL;DR: The concept of rules-data mutual enhancement, which can detect the ultimate relationship between data and rules with the utilization of two types of information resources: databases and knowledge bases, is introduced.
Abstract: This paper presents a new method of rules detecting about exploring hidden rules from a given knowledge base. The expended rule set forms a lattice that can be generated by the core. Not only does this method produce detecting rules, but also enhance data integrity and completeness. In addition, this paper introduces the concept of rules-data mutual enhancement, which can detect the ultimate relationship between data and rules with the utilization of two types of information resources: databases and knowledge bases. As a result, this paper makes initial endeavors to combine those two quite separated fields to promote the development of information technology.
TL;DR: How the current electronic commerce applications provide explicit and integrated support for complex agent interactions and present several virtual institutions where agents are engaged in multiple bilateral negotiations is highlighted.
Abstract: While there are already literature surveys upon agent-mediated electronic commerce applications, none have specifically tackled the issue from an interaction perspective or looked at how the control is distributed among the agents. This state-of-the-art survey focuses on how agent interactions are handled. First, it deeply looks at how methods for enforcing the actions taken by agents have been dealt with, namely protocols, negotiation and auction. Second, it defines the various types of communication languages used in multi-agent market architectures. The three main alternatives are KQML, ACL and FLBC. A comparison is then made between them and shows how much they suite their purpose. Third, this paper highlights how the current electronic commerce applications provide explicit and integrated support for complex agent interactions and present several virtual institutions where agents are engaged in multiple bilateral negotiations. Finally, it discusses some related research perspectives and identify some limitations.
Journal Article•
TL;DR: The results suggest that proper design of incentive structures and embedding these incentives into Group Decision Support System (GDSS) features is critical to the successful utilization of GDSS for virtual groups engaged in a mixed-motive negotiation task.
Abstract: In today's virtual organizations, the type of communication mode that individuals use to interact with others may influence their decision outcomes and the information they exchange with others. The information that the individuals exchange is often influenced by their incentive structures, the type of communication mode, and the nature of the authority structure in their group. This paper reports the results of a set of experiments in which the groups work on a mixed-motive negotiation task under different types of incentive structures, authority structures, and communication modes. The experimental study consists of a 2 × 2 × 2 factorial design. The results suggest that proper design of incentive structures and embedding these incentives into Group Decision Support System (GDSS) features is critical to the successful utilization of GDSS for virtual groups engaged in a mixed-motive negotiation task.
TL;DR: An architectural design is proposed that meets the needs of virtual organizations as far as any conflicts can be solved on an individual basis and Negotiating agents that obey the preferences of their principals are considered as a means to search for compromise solutions.
Abstract: Real and virtual organizations are conducted by preferences, goals, and policies. Since virtual organizations have a two-level structure the preferences of these different levels are a source of conflicting interests. Also, management policies itself are subject to change, especially in a virtual setting. Therefore, agility has superior value, causing the need to develop business information and planning systems that may be adapted easily. Both mentioned aspects of business planning software, namely, ease adaptation on one side and the integration of individual preferences and goals on the other are treated in the paper. An architectural design is proposed that meets the needs of virtual organizations as far as any conflicts can be solved on an individual basis. This concept is further elaborated. Negotiating agents that obey the preferences of their principals are considered as a means to search for compromise solutions. An extensive example derived from hospital management illustrates the concept.