In this paper we examine the issue of using granular computing (GrC) for intelligent query answering involving aggregate data, a topic related to applications of data mining. We point out that there are at least two general directions to deal with such a task, and examine these two directions by studying how to discover rules as intensional answers at the right levels of data granulation; and how to generate aggregate answers for queries at the right levels of data granulation.
In this paper we examine the problem of aggregate data mining in the context of granular computing (GrC). The motivation of this study is that although currently several GrC perspectives on data mining exist, the issue of mining on aggregate data still deserves more attention. We discuss the semantics of aggregate data mining and define aggregate data mining from a GrC perspective. Following a recently proposed framework of viewing granules as concepts, the problem of aggregate data mining is defined as how to infer the intension of data at granulation levels from given extensions at the level of base granules.
Recently integration of OLAP and data mining has drawn much attention, but there is still a lack of concrete result. In this paper we explore one aspect of such integration, called influential association rule mining. We first introduce the basic idea of association rule mining, and the two approaches, called IARM (Influential Association Rule Mining) and its improvement, IARMBM (Influential Association Rule Mining with BitMap) are briefly described. In addition, related experiments and comparisons are also reported.
Query processing in a mobile environment raises the question as to the manner in which to execute queries so that communication costs and response time are minimized. This paper investigates this issue and proposes a strategy for query optimization. We present several existing approaches of query processing in a mobile environment, establish a cost model to analyze the cost factors involved in each of the existing approaches and use this model to compare advantages and disadvantages of each approach. A new approach, entitled HERE THERE SQL, is then proposed to determine which of three pre-existing traditional approaches is the most efficient for any given circumstance. The basic motivation and the basic idea of this research are presented, and experimental results are sketched.
Mining association rules in transaction databases has received much attention in the field of data mining. Although progress has been made on techniques of mining association rules, the results often only indicate the mutual correlative relationships among the frequent items, paying no attention to the directional, or causal relations. For example, when a data set indicates an association between items A and B, it is often not clear whether the access of A caused the access of B, or the converse. In real world applications, however, knowing such causal relations is extremely useful for decision support. People would not only be interested in the facts that A and B are related, but also in the possible sequences and directions among the items. Mining transaction databases for this kind of knowledge offers the potential for deep analysis of business situations and finding strategies of operation. In this paper, we employ a Bayesian approach to mining causal relations from frequent itemsets. The results of our research include two algorithms based on Bayesian statistics model: a serial and diverging connection discovery algorithm (SDCD) and a converging connection discovery algorithm (CCD). Experimental results indicate that the performance of the algorithms is scalable.
Knowledge acquisition is important and machine learning techniques can be used to achieve automated knowledge acquisition. This article examines how knowledge acquisition can be assisted by programming using CLIPS (an acronym for C Language Integrated Production System). A machine learning preprocessor has been developed for the CLIPS environment, so that the CLIPS rule-base can be expanded by adding rules generated through machine learning techniques. The paper also shows how knowledge updating can be supported in the CLIPS environment itself. Operational engineering knowledge is captured in a data structure called a decision tree, and its structure can be updated when new knowledge is acquired. In addition, some advanced features are also briefly discussed, including using COOL (the CLIPS object-oriented language) for knowledge acquisition in a software product recommendation system, as well as the design of a self-evolving knowledge-acquisition tool.
In this article idea processors are studied as creativity supporting systems. Two related reviews are presented. The first review examines some aspects of idea processors and compares them with some existing work from AI research community. The second review is concerned with using retrospective analysis for technical invention (particularly the invention of artifacts). Retrospective analysis not only opens new opportunities for knowledge-based approaches in idea generation, but also offers new mechanisms to deal with some problems faced by idea processors using object-oriented features. To illustrate our new approach of idea generation, we also provide a brief sketch of our project which uses a Prolog program to implement some key ideas presented in this article.
Knowledge discovery in databases (KDD) and data mining have good potential for better management of organizations. In this article we describe a conceptual model for user-guided data mining, and a methodology for query construction based on this model. The methodology allows the user to express what kind of knowledge is to be discovered, thus incorporating user intention to alleviate the overabundance problem which has hampered the development of data mining. A user starts from the goal at the top-most level, and refines queries under the guide of a causal network. The process of query construction is illustrated by examples.
Idea processors, as a kind of software widely used in the business world, have not received much attention from academia. In this paper we provide an overview of the current status of idea processors. We start from the foundations of idea processors, pointing out their roots in brainstorming techniques. By examining several experimental systems and commercial products, we further discuss how idea processors work, their nature, and their typical architecture. We also summarize some research work related to idea processors, as well as relationships between idea processors and studies of computational creativity in artificial intelligence. Other related issues, such as group decision support systems and evaluation methods, are also briefly examined.
A data warehousing approach for recommender sys- tems is proposed. We sketch an architecture for integrated OLAP and data mining in data waxe- housing environments, and argue why this archi- tecture can be extended for building recommender systems. Since producing recommendations can be considered as conceptual query answering, the re- lationship between conceptual query answering and intensional answers is also briefly examined.
An expert critic (or an expert critiquing system) is a computer program that critiques human-generated solutions. So far, ethics has not been incorporated in critiquing systems. We investigate the important issue of incorporating ethical issue testing into expert critic systems. To illustrate the need for such a study, our discussion consists of ethical issue testing in physicians' treatment plans for cancer patients. Three ethical issues are identified, and one of them, namely, the Guinea pig treatment ethics, is discussed in detail. An experimental prototype system (ETHICS1) has been developed for this purpose. The structure of this system along with its implementation is described. We compare our study with existing expert critiquing systems and point out some implications of building critics for ethical issue testing.
Integration of symbolism and connectionism is an important issue for building intelligent systems. It is also a challenging issue, due to the different medium of representations required for these two paradigms. In this paper the author explores the synergy of connectionism and expert systems (or, more generally, knowledge-based systems) through an expanding approach; namely, the architecture of classical expert systems is expanded by attaching a connectionist learning assistant. The connectionist learning assistant is able to learn by observing user behavior and capturing the results of observations; these results are used in conjunction with the inference engine so that flexible inference control can be achieved and the adaptability of expert systems can be enhanced. In this article, the author explains why the expanding approach is needed, why flexible control of the inference engine is important, and how this can be realized by coupling a connectionist learning component with the inference engine. These considerations result in an expanded knowledge-based system architecture, which features a flexible searching controller attached on an inference engine, a connectionist learning assistant, and a user modeller
We examine the relationship between systems and their users from the knowledge discovery perspective. Recently knowledge discovery in databases has made important progress, but it may also bring some potential problems to database design, such as issues related to database security, because an unauthorised user may derive highly sensitive knowledge from unclassified data. In this paper we point out that there is a need for a comprehensive study on knowledge discovery in human-computer symbiosis. Borrowing terms from algorithm design and artificial intelligence literature, we propose a notion called database-user adversarial partnership. We point out that this notion is general enough to cover various knowledge discovery and security of issues related to databases and their users. Furthermore, we point out the notion of database-user adversarial partnership can be further generalised into system-user adversarial partnership. Opportunities provided by knowledge discovery techniques and potential social implications are also discussed.
In a discussion concerning the future of the library, Minsky and Feigenbaum endorsed the idea for books to ‘talk to each other’. Creating knowledge or doing inference by connecting related documents is an important aspect of undiscovered public knowledge , an important concept initiated by Swanson. However, so far few computer models exist in this regard. In this paper, we describe such a model which connects documents. We first explain the meaning of integration of scientific texts by connecting documents. Our fundamental idea is then illustrated by an example and further described by a model. Some features of this model are discussed. A brief comparison with some other computer models is also provided.
The role of user responsibility in decision support systems (DSS) has been omitted in recent studies of some important issues. One such important issue is exception handling. In this short note, the role of user responsibility in exception handling is examined. The importance of user responsibility is emphasized, a possible approach is outlined which supports user responsibility in exception handling through the consideration of mental models.
In this paper, we discuss the integrated use of expert systems built from expert systems shells. We explain that this integration can be conceptually viewed as expert systems integration at the K-tree level. The relationship between our work and other work in distributed artificial intelligence is also examined.
The contribution of systems theory is reexamined from an organizational view. Its influence on artificial intelligence (and on computer science) is examined through the examples of analogical problem solving and cooperative distributed problem solving.
Although the study of analogy is a common interest for researchers in general systems theory and in artificial intelligence (AI), the research disciplines in these two areas are quite different. In this paper we provide a brief comparison study for these two fields: analogy as the study of systems and analogy as the study of intelligence. We show that these two fields of study are complementary. We also propose the use of logic specification for systems description. A study of analogy that combines perspectives of general systems theory and AI thus becomes evident.
Case-based reasoning is now a well-explored method for intelligent problem solving. It has great potential for engineering applications, particularly for those related to reuse and redesign. However, to make it practically useful, an important technical problem to be solved is how to realize the integrated use of existing case libraries. In this paper, we first give an overview of case-based reasoning, then outline our methodology which leads to a distributed case library in a network environment. We explain how the indexing and retrieval techniques as used in “conventional” case libraries can be extended. An experimental system is described, and the relationship between our work and an object-oriented paradigm, as well as the relationship between our work and distributed artificial intelligence are also briefly examined.
