Expert system developer testing

Expert System Developer Testing Procedures. An expert system developer can generate his/her own code for testing various elements of the ES code without going to fully automated self-contained procedures. Borrowing from the conventional software side, three... 

The testing phase is important in expert system development. The practical applicability of the expert system will largely depend on this phase. Testing expert systems is different from normal software engineering in a number of ways. First, it is difficult to test exhaustively the full code and all possible paths the reasoning process may follow. Secondly, the nature of expert systems poses some typical problems. Due to their heuristic nature the correctness of the results cannot be easily verified. A certain degree of errors may be acceptable and, moreover, an... 

This chapter provides an overview of expert-system verification and validation (V V) techniques. Several methods are presented. First, many of the conventional software V V techniques such as requirements analysis and unit testing can be applied to expert-system development. Second, an expert-system developer can use automated tools to test rule consistency and structure. A more viable alternative, however, is for the developer to create his own set of consistency and completeness tests. Finally, a developer should rely on qualitative judgment to determine the validity of a knowledge base. This judgment could include expert opinion as well as specialized tests designed to determine knowledge-base certification. The chapter suggests that methods should be combined into an optimal mix in order to best undertake V V. 

Validation involves determining that the system performs with a reasonable level of accuracy. Validation is accomplished through test and evaluation of ES software and integrated hardware. Validation thus ensures that the capabilities that have been specified in the ES requirements have been exercised and meet levels acceptable to the user. Thus, without a true V V methodology, much time is lost in the evolutionary expert system development process. 

A designer should consult with experts to interpret data from chemical compatibility tests. To meet this need, U.S. EPA developed a software system called Flexible Membrane Liner Advisory Expert System (FLEX) to assist in evaluating test data. FLEX is an expert system that is based on data from many chemical compatibility tests and contains interpretations from experts in the field. 

Most expert systems are built incrementally — a small prototype system is built and tested before further rules are added. This reduces the risk that the full system will contain errors that are not found until practical completion, when tracking down their origin can be difficult, and also gives the expert an early chance to test the look-and-feel of the system, so that any changes required in its appearance can be incorporated before development has proceeded too far. 

A number of software packages or expert systems for ruggedness testing has been developed. RES (commercialized under the name Shaiker ) is an expert system created by Van Leeuwen et al. [4,23] and has been validated and evaluated [42,43]. It uses fractional factorial and Plackett-Burman designs and allows to test the factors at two or three levels. The interpretation criteria used here are the predefined values (see Section 3.4.8). 

Expert systems can also be used off-line as advisory systems. If some leeway is left in the process specifications, operators can even be instructed to make simple variations in the process. Hajicek developed an expert system for troubleshooting injection molding, which was designed to be used by operators to adjust controls as needed whenever part quality suffered [20], These expert systems do not necessarily plan an entire cycle. As we will see in the section on in-process control, however, they can be used to do so. There are even specifications that include rules for altering the cure cycle in response to the results of quality acceptance tests such as flow and volatiles, a primitive set of heuristics. 

The system operational test cases may be developed with the collaboration of the application support personnel, the application expert, or a system user. 

Current generally applicable biodegradation models focus on the estimation of readily and nonreadily biodegradability in screening tests. This is because most experimental data are from such tests (e.g., MITI-I). There are far fewer data that are both quantitative and environmentally relevant (i.e., measured half-lives or rate constants). However, individual transformations and pathways are well documented in the literature. This allows for development of explicitly mechanistic models, making use of established group-contribution approaches, hierarchic rule-based expert systems, and probabilistic evaluation of possible transformation pathways. 

Ames BN, Durston WE, Yamasaki E, Lee FD (1973) Carcinogens are mutagens a simple test system combining liver homogenates for activation and bacteria for detection. Proc Natl Acad Sci USA 70 2281-2285 Barratt MD, Castell JV, Miranda MA, Langowski JJ (2000) Development of an expert system rulebase for the prospective identification of photoallergens. J Photochem Photobiol B 58 54-61... 

Taking any test can be tough. But, don t let the written test scare you If you prepare ahead of time, you can achieve a top score. The LeamingExpress Test Preparation System, developed exclusively for LeamingExpress by leading test experts, gives you the discipline and attitude you need to be a winner. 

System improvements may improve the quality of prescribing by standardizing to an expert level. Morris (51) described the development, testing, and use of computerized protocols for management of intravenous fluid and hemodynamic factors in patients with acute respiratory distress syndrome. Evans et al. (52) used a computerized anti-infectives management... 

Metabolism databases and several predictive software packages also have been reviewed.A knowledge-based expert system for the prediction of phase I and II biotransformations called META was developed on the VAX/VMS platform,META contains over 750 biotransformations based on substrnctnres and qnantnm mechanical calculations that gave excellent predictions on test data and have been optimized using a genetic algorithm to perform better than human experts.However, when used to predict the metabolism of 42 polycyclic aromatic hydrocarbons (PAHs), META overpredicted 29 and missed 8 of 72 experimentally observed epoxidations, and missed 27 of 49 experimentally observed hydroxylations. ... 

Analytic evaluation methods can be used early in the development process, well before there are users or prototypes available for empirical tests. Furthermore, it is often less expensive than making studies with users. Examples of analytic methods are heuristic evaluation, cognitive walkthroughs, usability-expert reviews, group design reviews [864, 1025]. A hazard of analytic evaluation is that system developers or software designers may feel that they are being evaluated [941]. 

---