One widely used knowledge-discovery technique is a decision-tree inducer that generates classifiers in the form of a single decision tree. As the number of prespecified decision-outcome classes increases, however, the trees so generated often become overly complex with regard to the number of leaves and nodes, and the classification accuracy consequently drops. In contrast, the multi-decision-tree induction (MDTI) approach, which constructs different decision trees for different decision-outcome classes, may reduce rule cardinality, and improve both rule conciseness and classification accuracy over a traditional single-decision tree inducer. This paper analytically and empirically compares the two techniques based on these measures. The analysis and results show that, in some situations, MDTI outperforms the traditional approach in terms of cardinality, conciseness, and classification accuracy of the acquired knowledge structures.

The rapid advancement of Internet technologies enables more and more educational institutes, companies, and government agencies to provide services, namely online services, through web portals. With hundreds of online services provided through a web portal, it is critical to design web portals, namely service portals, through which online services can be easily accessed by their consumers. This article addresses this critical issue from the perspective of service selection, that is, how to select a small number of service-links (i.e., hyperlinks pointing to online services) to be featured in the homepage of a service portal such that users can be directed to find the online services they seek most effectively. We propose a mathematically formulated metric to measure the effectiveness of the selected service-links in directing users to locate their desired online services and formally define the service selection problem. A solution method, ServiceFinder, is then proposed. Using real-world data obtained from the Utah State Government service portal, we show that ServiceFinder outperforms both the current practice of service selection and previous algorithms for adaptive website design. We also show that the performance of ServiceFinder is close to that of the optimal solution resulting from exhaustive search.

Web search engines typically provide search results without considering user interests or context. We propose a personalized search approach that can easily extend a conventional search engine on the client side. Our mapping framework automatically maps a set of known user interests onto a group of categories in the Open Directory Project (ODP) and takes advantage of manually edited data available in ODP for training text classifiers that correspond to, and therefore categorize and personalize search results according to user interests. In two sets of controlled experiments, we compare our personalized categorization system (PCAT) with a list interface system (LIST) that mimics a typical search engine and with a nonpersonalized categorization system (CAT). In both experiments, we analyze system performances on the basis of the type of task and query length. We find that PCAT is preferable to LIST for information gathering types of tasks and for searches with short queries, and PCAT outperforms CAT in both information gathering and finding types of tasks, and for searches associated with free-form queries. From the subjects' answers to a questionnaire, we find that PCAT is perceived as a system that can find relevant Web pages quicker and easier than LIST and CAT.

The need for network storage has been increasing rapidly owing to the widespread use of the Internet in organizations and the shortage of local storage space due to the increasing size of applications and databases. Proliferation of network storage systems entails a significant increase in the number of storage objects (e.g., files) stored, the number of concurrent clients, and the size and number of storage objects transferred between the systems and their clients. Performance (e.g., client-perceived latency) of these systems becomes a major concern. Previous research has explored techniques for scaling up the number of storage servers involved to enhance the performance of network storage systems. However, adding servers to improve system performance is an expensive solution. Moreover, for a WAN-based network storage system, the bottleneck for its performance improvement typically is not caused by the load of storage servers, but by the network traffic between clients and storage servers. This paper introduces an Internet-based network storage system named NetShark and proposes a caching-based performance-enhancement solution for such a system. The proposed performance-enhancement solution is validated using a simulation.

Abstract A large number of studies have investigated the transaction log of general‐purpose search engines such as Excite and AltaVista, but few studies have reported on the analysis of search logs for search engines that are limited to particular Web sites, namely, Web site search engines. In this article, we report our research on analyzing the search logs of the search engine of the Utah state government Web site. Our results show that some statistics, such as the number of search terms per query, of Web users are the same for general‐purpose search engines and Web site search engines, but others, such as the search topics and the terms used, are considerably different. Possible reasons for the differences include the focused domain of Web site search engines and users' different information needs. The findings are useful for Web site developers to improve the performance of their services provided on the Web and for researchers to conduct further research in this area. The analysis also can be applied in e‐government research by investigating how information should be delivered to users in government Web sites.

The expanded role of radiology in clinical medicine and its emerging digital practice have made patient-image management a growing concern for health-care organizations. A fundamental aspect of patient-image management is to provide a radiologist with convenient access to prior images relevant to his or her reading of a recently taken radiological examination. For confirmation or evaluation purposes, radiologists often reference relevant prior images of the same patient when interpreting the images of a current examination. To alleviate the time and physical requirements on radiologists, many health-care organizations have taken a prefetching strategy for meeting their patient-image reference needs. Radiologists' patient-image reference knowledge understandably may exhibit subtle individual variations and dynamically evolves over time, thus making the artificial intelligence-based inductive learning approach appealing. Central to patient-image prefetching is a knowledge base of which knowledge elements need continual update and individual customization. In this study, we extended a decision rule induction technique (i.e., CN2 algorithm) to address the challenging characteristics of the targeted learning. We experimentally evaluated the extended algorithm using the learning performances achieved by backpropagation neural network as benchmarks. Overall, our evaluation results suggest that the extended algorithm exhibited satisfactory learning effectiveness and, at the same time, showed desirable noise tolerance, immunity to missing data, and robustness in relation to limited training data.

As the size and complexity of Web sites expands dramatically, it has become increasingly challenging to design Web sites where Web surfers can easily find the information they seek. In this article, we address the design of the portal page of a Web site, which serves as the homepage of a Web site or a default Web portal. We define an important research problem---hyperlink selection: selecting from a large set of hyperlinks in a given Web site, a limited number of hyperlinks for inclusion in a portal page. The objective of hyperlink selection is to maximize the efficiency, effectiveness, and usage of a Web site's portal page. We propose a heuristic approach to hyperlink selection, LinkSelector, which is based on relationships among hyperlinks---structural relationships that can be extracted from an existing Web site and access relationships that can be discovered from a Web log. We compared the performance of LinkSelector with that of the current practice of hyperlink selection (i.e., manual hyperlink selection by domain experts), using data obtained from the University of Arizona Web site. Results showed that LinkSelector outperformed the current manual selection method.

When performing primary reading on a newly taken radiological examination, a radiologist often needs to reference relevant prior images of the same patient for confirmation or comparison purposes. Support of such image references is of clinical importance and may have significant effects on radiologists' examination reading efficiency, service quality, and work satisfaction. To effectively support such image reference needs, we proposed and developed a knowledge-based patient image pre-fetching system, addressing several challenging requirements of the application that include representation and learning of image reference heuristics and management of data-intensive knowledge inferencing. Moreover, the system demands an extensible and maintainable architecture design capable of effectively adapting to a dynamic environment characterized by heterogeneous and autonomous data source systems. In this paper, we developed a synthesized object-oriented entity- relationship model, a conceptual model appropriate for representing radiologists' prior image reference heuristics that are heuristic oriented and data intensive. We detailed the system architecture and design of the knowledge-based patient image pre-fetching system. Our architecture design is based on a client-mediator-server framework, capable of coping with a dynamic environment characterized by distributed, heterogeneous, and highly autonomous data source systems. To adapt to changes in radiologists' patient prior image reference heuristics, ID3-based multidecision-tree induction and CN2-based multidecision induction learning techniques were developed and evaluated. Experimentally, we examined effects of the pre-fetching system we created on radiologists' examination readings. Preliminary results show that the knowledge-based patient image pre-fetching system more accurately supports radiologists' patient prior image reference needs than the current practice adopted at the study site and that radiologists may become more efficient, consultatively effective, and better satisfied when supported by the pre-fetching system than when relying on the study site's pre-fetching practice.

This contingency study was undertaken to increase understanding of the way in which management support and communication characteristics influence teleworkers' use of communication media. Complexity of communication tasks, communication urgency, and perceived reliability of a medium comprise variables of communication characteristics. These variables represent theoretical dimensions of media richness: perceived task requirement, perceived media characteristics, and attitude toward communication media. Information richness and social influeuce provided the theoretical basis of this study. The study focused on two general-purpose communication media, e-mail and the telephone, and examined how the above factors are used to determine the choice of media. Research data were collected from the formal telework program operated by Fujitsu Kyushu Communication Systems Limited (QCS) in Japan. We differentiate our study from previous studies of media choice by focusing on remotely distributed teleworkers, using clearly defined work groups that permit investigation of collective behaviors, and analyzing data captured during actual use of communication media. Analysis of data indicates that factors of communication characteristics, especially communication urgency and perceived reliability of a communication medium, do influence the media use of teleworkers. Management support, as a direct social influence as well as a moderating force that overcomes the constraint of media richness, greatly improved the adoption and effective use of e-mail.

Reviews and compares five medical information exchange standards: the HL7 standard, the American Society of Testing and Materials Medical standard (E1238), the Medical Information Bus standard (IEEE P1073), the Medical Data Interchange Standard (IEEE P1157), and the American College of Radiology and National Electrical Manufacturers Association (ACR-NEMA) standard. Each standard has its primary goal stated, and its relationship to networking protocols explored. Our conclusion is that the standards that proliferate most will be the ones which utilize a practical networking approach. These standards will focus their standardization efforts outside the communication network solution space (i.e. with minimal definition of protocol layers).>

Electronic Mail Systems (EMSs) have been a boon to organizational productivity, but these systems may not receive greater acceptance due in part to their general lack of robust system features. Particularly, the intractable problem of effective electronic mail dissemination management has largely eluded solution. This paper first provides a brief survey of several existing intelligent EMSs which seek to address the conundrum of mail management and then proposes an alternative solution which takes advantage of the unique characteristics of a knowledge base/database coupling to facilitate effective information sharing among members of an organization. This approach improves on prior Knowledge-Based Mail System (KMS) designs by (1) focusing on message dissemination management rather than on message receipt management, (2) separating the relatively static knowledge of message routing from the relatively dynamic knowledge of the organizational membership, and (3) incorporating a more comprehensive view of the semantic constraints involved in configuring distribution lists of message receivers through the use of the Structured Object Model (SOM) Methodology. The design concepts are illustrated by a prototype KMS, the Message Dissemination System (MDS).

The objective is to develop a structured object-oriented modeling and design methodology for coupled knowledge-base/database (KB/DB) systems by exploring the useful principles and features of object-oriented modeling and software development techniques. The methodology uses a synthesize object-oriented entity-relationship model for representing the knowledge and the embedded data semantics involved in coupled KB/DB systems. An associated design procedure is presented. This methodology improves on existing coupled KB/DB design methods because of its well-defined constructs that deal with various forms of knowledge involved in data processing, knowledge-based problem solving and object-oriented reasoning.>

Expert systems are powerful tools that serve as adjuncts to decision making and have found wide applicability in a wide variety of areas. Integrating expert systems with group decision support systems has the potential to enhance the quality and effeciency of group communication, negotiation, and collaborative work. This paper examines possible synergies between the two technologies and provides a survey of current partially-integrated systems. Finally, a prototype design of a highly-integrated system is described with directions for further research.

A well-designed database that contains relevant, nonredundant, and consistent data is rarely achieved in practice. One major reason for this is seen to be the lack of effective support for logical database design. In view of this, the structured entity model (SEM) method, an object-oriented method which facilitates the logical design phases of database system development was developed. Because, unlike most methods, SEM does not require designers to understand normalization theory in order to design a logical database, an end user (or a novice designer) can successfully complete logical database design using the method.>