法則庫系統之評量

法則庫系統已被廣為應用於各種問題領域中。而法則庫系統發展之管理則迫切需要適切的評量方法。本研究的目的乃是提出評量法則庫系統複雜度之方法，並藉以反應出其發展所需人力。法則庫系統之評量包刮事實、法則、與推理等。因為本研究所提出的評量方法具有表達性與擴充性，所以可應用於法則庫系統發展之管理。「專家系統諮詢顧問」則被用已闡述這些評量方法的使用。

關鍵字

複雜度；事實評量法；推理評量法；推理比率；法則庫系統

並列摘要

The objective of this study is to propose measurements for the complexities of rule based systems so as to indicate development efforts. The measurements of rule based systems include those for facts, rules, and inference relationships. Fact counts are used to measure the amounts of facts. Measurements for rules consist of volume, conditions and actions, and number of variables and wildcards. Inference measurements reflect inference relationships among rules, including distribution of rule relationships, inference ratio, and inference count. Empirical analyses are performed on Expert System Consultant (an expert system for selecting ESBTs) to illustrate the usage of these measurements. From the analyses, it is demonstrated that the measurements possess desired characteristics such as expressiveness and extendibility, and can be used for managing the development of rule based systems.

並列關鍵字

Complexities ； fact count ； inference count ； inference ratio ； rule based system

參考文獻

Brooks, F. P.(1975).Mythical Man-Month: Essays on software engineering.Reading, MA:Addison-Wesley.

Google Scholar

Brownston, L.,Farrell, E.,Kant, E.,Martin, N.(1985).Programming Expert Systems in OPS5: An Introduction to Rule-Based Programming.Reading, MA:Addison-Wesley.