Key factors in knowledge acquisition

Abstract

In recent years there has been a tremendous increase in the development of Expert Systems (ESs) in organizations. This increased development is straining the already limited supply of qualified ES developers. These ES developers have come to be know as Knowledge Engineers (KEs), and their job as Knowledge Engineering. The process of Knowledge Engineering is divided into two tasks: Knowledge Acquisition (KA) and ES construction. KA has been defined as "The process of extracting, structuring, and organizing Knowledge from several sources, usually human experts, so it can be used in a program." (Waterman, 1986; p. 392) This process of KA has been identified as the "bottleneck" that currently constrains the development of ESs.This paper summarizes what is known about the KA process in an effort to identify what the key factors are that influence the success of the KA phase of the Knowledge Engineering process. Due to the similarities that exist between ESs and traditional systems development, the literature that pertains to traditional Information Requirements Determination and to Systems Analysts will be utilized to guide this exploration. Case study reports of actual ES development projects and the practitioner literature from this highly applied field will also be referenced. A model of the knowledge Engineering process has been developed and will be used to help determine and discuss the key factors that influence the KA process.Five key factors have been identified and will be discussed in detail. These factors are: the attributes of the participants in the ES development process; elicitation techniques utilized in the process; the development of external representations; representation selection problems; and the verification of the ES by the Domain Expert who participates in its continued development via the user-system interface. These factors are presented, key issues identified, and research questions suggested for each area.It is hoped that the analysis of the key factors in KA will lead to the identification of the skills and techniques necessary to successfully perform the KA process. Once these skills have been identified, training programs can be developed to help reduce the shortage of qualified KEs and, ultimately, facilitate the increased development of ESs in organizations.

Scientific Knowledge Engineering: a conceptual delineation and overview of the state of the art

“…Even though the modelling of scientific results into knowledge bases is central to SKE, it is important to see that the proposition of SKE does not have to be restricted to this aspect. This is because the successful development of knowledge-based systems involve many other factors (Freiling et al ., 1985; Dibble & Bostrom, 1987; Fellers, 1987; Rook & Croghan, 1989; Motta et al ., 1990; Plant, 1991; Studer et al ., 1998; Schreiber, 2000). This extensive technical literature on KE processes and procedures form a solid basis upon which SKE can be systematized to guide one through most of the relevant technical decisions made in specifying and building scientific computational infrastructures.…”

“…But as mentioned earlier, they will be left to future SKE initiatives. Returning our attention to the management level, the following section presents a management-level process based on several works which proposed some kind of process organization for KE projects (Freiling et al ., 1985; Dibble & Bostrom, 1987; Fellers, 1987; Rook & Croghan, 1989; Motta et al ., 1990; Plant, 1991; Studer et al ., 1998; Schreiber, 2000). It should be noticed that, on the management level, the focus will be more on what should be done than on how it should be done.…”

“…Furthermore, it offers the possibility to explore the domain area and to conduct some kind of validation of concepts and relationships structures commonly found in the defined scope. This progressive specification starting from the conceptual level up to the implementation level appears with different names in the technical literature such as shared and external representations in Fellers (1987), micro- and macro-knowledge in Rook and Croghan (1989), and primary representations and domain specifications in Plant (1991). As a result of this incremental specification and given the relative lower commitment of this step compared with the following steps, this point can be an appropriate moment to consider if a new iteration is necessary for better familiarization with the domain or for scope refinement.…”

Selection of knowledge acquisition techniques based upon the problem domain characteristics of production and operations management expert systems

“…Previous researchers have recognized the need for sound empirical research to compare the effectiveness and efficiency of knowledge acquisition tools and methods. Fellers (1987) concluded that more research was needed to answer the following.…”

Case research on knowledge acquisition: observations and lessons

“…Hart (1986) supported this approach. Similarly, Fellers (1987) viewed knowledge engineering to consist of two phases: knowledge acquisition and expert system construction. The former process involves "one or more knowledge engineers interacting with one or more domain experts (DEs)", and the objective of this interaction is to develop a "shared representation" or model of the DE's problem-solving process.…”