Impediments to Exploratory Data Mining Success

This chapter discusses impediments to exploratory data mining success. These impediments were identified based on anecdotal observations from multiple projects either reviewed or undertaken by the author and are classified into four main areas: data quality; lack of secondary or supporting data; insufficient analysis manpower; lack of openness to new results. Each is explained, and recommendations are made to prevent the impediment from interfering with the organization’s data mining efforts. The intent of the chapter is to provide an organization with a structure to anticipate these problems and to prevent the occurrence of these problems.

An Intelligent Support System Integrating Data Mining and Online Analytical Processing

Intelligent decision support requires flexible, knowledge-driven analysis of data to solve complex decision problems faced by contemporary decision makers. Recently, online analytical processing (OLAP) and data mining have received much attention from researchers and practitioner alike, as components of an intelligent decision support environment. Little that has been done in developing models to integrate the capabilities of data mining and online analytical processing to provide a systematic model for intelligent decision making that allows users to examine multiple views of the data that are generated using knowledge about the environment and the decision problem domain. This paper presents an integrated model in which data mining and online analytical processing complement each other to support intelligent decision making for data rich environments. The integrated approach models system behaviors that are of interest to decision makers; predicts the occurrence of such behaviors; provides support to explain the occurrence of such behaviors and supports decision making to identify a course of action to manage these behaviors.

The Role of Data Mining in Organizational Cognition

Data mining has quickly emerged as a tool that can allow organizations to exploit their information assets. In this chapter, we suggest how this tool can be used to support strategic decision-making. Starting with an interpretive perspective of strategy formulation, we discuss the role of beliefs in the decision-making process. Referred to as Micro-Theories (MTs), these beliefs generally concern some assumption regarding the organization’s task environment, such as sales increasing in a certain segment or customers preferring a certain product. The strategic role for data mining, referred to as Organizational Data Mining (ODM) is then to provide validation for these beliefs. We suggest a four-step process for identifying and verifying MTs and illustrate this with a hypothetical example of a bank. Implications and future trends in ODM are discussed. Ultimately results of data mining should be integrated with strategic support systems and knowledge management systems.

Gaining Strategic Advantage Through Bibliomining

Library and information services in corporations, schools, universities and communities capture information about their users, circulation history, resources in the collection and search patterns (Koenig, 1985). Unfortunately, few libraries have taken advantage of these data as a way to improve customer service, manage acquisition budgets or influence strategic decision making about uses of information in their organizations. In this chapter, we present a global view of the data generated in libraries, and the variety of decisions that those data can inform. We describe ways in which library and information managers can use data mining in their libraries, i.e., bibliomining, to understand patterns of behavior among library users and staff members and patterns of information resource use throughout the institution. The chapter examines data sources and possible applications of data mining techniques in the library.

Empowering Modern Managers

This chapter discusses recent advances in the use of agent technology in Decision Support Systems (DSSs) and introduces a model for an agent-based DSS. The chapter analyzes the modern requirements for the nature of decision support and argues in favor of adopting active situated paradigm as the basis for building DSS. The benefits of agent technology are highlighted in relation to the desired features of DSS and the past research in this direction is reviewed and systematically categorized. The description of an agent-based DSS elaborating on the architecture of the system and the potential use of data mining techniques is then introduced. The approach is illustrated with an agent-based DSS for investment decisions. The chapter informs the readers about the state of art in agent-based DSS, and provides a framework that can be used as a reference model in future research in the area.

Knowledge Mining in DSS Model Analysis

The three stages of mathematical modeling include model formulation, solution and analysis. To date, the primary focus of model-based decision support systems (DSS), in general, and Management Science/Operations Research (MS/OR), specifically, has been on model formulation and solution. In fact, with a few notable exceptions, computer-assisted model analysis has been ignored in both information systems (IS) and MS/OR literature (Swanson & Ramiller, 1993). This lack of attention to model analysis is especially noteworthy for two reasons. First, the primary bottleneck of modeling is in the analysis and interpretation of model results (Greenberg, 1993). Second, the basic purpose of DSS and mathematical modeling is insightful understanding of the modeled environment through insightful analysis (Geoffrion, 1976; Steiger, 1998). Developing insight into the complex decision-making environment is ultimately a process of discovery, finding trends and surprising behaviors and comparing the behavior of the model to what is expected or observed in the real system (Jones, 1992). Thus, insightful understanding often entails the inductive analysis of several (if not many) model instances (i.e., what-if cases), each of which has one or more different values for input parameters in an attempt to understand the associated changes in the modeled output.

Mining Meaning

This chapter focuses on the pattern detection and extraction step in text data commonly called text data mining. I examine some of the literature on natural language processing and propose a method of recovering value from the text of virtual group discussions based on methods derived from the communication field. Then, I apply the method in a case using data from 216 different groups from a virtual group experiment. The results from the case show that higher performing groups are characterized by higher frequencies of acts of dominance and higher frequencies of terms concerning cognition, communication and praise. Higher performing groups were also characterized by lower frequencies of acts of equivalence and lower frequencies of leveling terms and numerical terms. Ways to use this knowledge to improve the groups’ performance are discussed.

Organic Knowledge Management for Web-Based Customer Service

This chapter introduces practical issues of information navigation and organizational knowledge management involved in delivering customer service via the Internet. An adaptive, organic approach is presented that addresses these issues. This approach relies on both a system architecture that embodies effective knowledge processes, and a knowledge base that is supplemented with meta-information acquired automatically through various data mining and artificial intelligence techniques. An application implementing this approach, RightNow eService Center, and the algorithms supporting it are described. Case studies of the use of eService Center by commercial, governmental and other types of organizations are presented and discussed. It is suggested that the organic approach is effective in a variety of information-providing settings beyond conventional customer service.

Data Mining in Franchise Organizations

Franchising has been used by businesses as a growth strategy. Based on the authors’ cumulative research and experience in the industry, this paper describes a comprehensive framework that describes both the franchise environment — from customer services to internal operations — and the pertinent data items in the system. The authors identify the most important aspects of a franchising business, the role of online analytical processing (OLAP) and data mining play and the data items that data mining should focus on to ensure its success.

Mining Message Board Content on the World Wide Web for Organizational Information

This chapter proposes a methodology to scan, analyze and classify the content of primarily text-based Web documents to aid an organization in gathering information. The representation and classification of the document is based on the popular vector space model and linear discriminant analysis, respectively. The methodology is developed and demonstrated using real chat room discussions about a publicly traded company collected over a 12-day period. The purpose of this chapter is to develop and demonstrate a methodology used to aid an organization in its environmental scanning efforts, in light of the vast quantities of information available via the Internet.