Rule-based expert system structure

Rules seemingly have the same format as IF.. THEN.. statements in any other conventional computer language. The major difference is that the latter statements are constructed to be executed sequentially and always in the same order, whereas expert system rules are meant as little independent pieces of knowledge. It is the task of the inference engine to recognize the applicable rules. This may be different in different situations. There is no preset order in which the rules must be executed. Clarity of the rule base is an essential characteristic because it must be possible to control and follow the system on reasoning errors. The structuring of rules into rule sets favours comprehensibility and allows a more efficient consultation of the system. Because of the natural resemblance to real expertise, rule-based expert systems are the most popular. Many of the earlier developed systems are pure rule-based systems. 

Expert systems have been defined as any formal systems , which make predictions about the toxicity of chemicals. All expert systems for the prediction of toxicity are built on experimental data and/or rules derived from such data (Dearden 2003). The expert systems can be further divided into two subclasses based on the method of generating rules. The one method is a knowledge- or rule-based expert system, for which experts/toxicologists create rules based on a list of structural features that have been related to a specified toxicity (Durham and Pearl 2001). An example of a typical knowledge- or rule-based system is DEREK, which will be described later. 

The IDD Advisor is a structured rule-based expert system comprised of several rule sets, the organization of which is illustrated in the context tree in Figure 1. Each node represents a separate rule base which, provides consultation on separate processes in the overall IDD process. 

In addition to starting material selection, CAESA carries out an analysis of the structural features of a molecule that give rise to synthetic difficulty or complexity using a rule-based expert system. This system identifies and quantifies the molecular complexity that results from the topology, the stereochemistry. 

Chemical metabolism can be described qualitatively or quantitatively. Many scientists can make qualitative predictions of the likely excretion products or blood plasma metabolites in mammals, or a particular animal including man, based on accumulated knowledge and experience. Such knowledge, in its raw form, generally consists of structure-metabolism relationships that are frequently expressible as qualitative structure-based rules that may be encoded into computer-based expert systems (see Chapter 9 for a full definition). Examples of such systems, in their more fully developed commercial forms, are discussed toward the end of this chapter. 

DfW is a knowledge-based expert system for predicting the toxicity of a chemical from its molecular structure (Judson et al., 2003). This system is composed of structural alerts, example compounds, and rules that may each contribute to the toxicity predictions. However, DfW is designed to aid in chemical carcinogenicity risk assessment, and hepatotoxicity is not the major toxicity prediction from this... 

In these frames all specific columns that are relevant for the reasoning process of the expert system can be described in a structured and comprehensive way. The frame-based and rule-based knowledge representation are both required to represent expertise in a natural way. Therefore, in most expert systems a combination of rule-based and frame-based knowledge representation is used. The rule base together with the factual and descriptive knowledge by means, of e.g., frames constitute the knowledge base of the expert system. 

Figure 13.11 Overview diagram of the NCTR Four-Phase approach for priority setting. In Phase I, chemicals with molecular weight < 94 or > 1000 or containing no ring structure will be rejected. In Phase II, three approaches (structural alerts, pharmacophores, and classification methods) that include a total of 11 models are used to make a qualitative activity prediction. In Phase III, a 3D QSAR/CoMFA model is used to make a more accurate quantitative activity prediction. In Phase IV, an expert system is expected to make a decision on priority setting based on a set of rules. Different phases are hierarchical different methods within each phase are complementary.

Zinke, S., Gerner, I., and Schlede, E., Evaluation of a rule base for identifying contact allergens by using a regulatory database comparison of data on chemicals notified in the European union with "structural alerts" used in the DEREK expert system, Alt. Lab. Anim. (ATLA), 30, 285-298, 2002. 

---