Organizing knowledge for tutoring fire loss prevention |
| |
Authors: | Daniel L. Schmoldt |
| |
Affiliation: | (1) Pacific Southwest Forest and Range Experiment Station, USDA Forest Service, 4955 Canyon Crest Drive, 92507 Riverside, California, USA |
| |
Abstract: | The San Bernardino National Forest in southern California has recently developed a systematic approach to wildfire prevention planning. However, a comprehensive document or other mechanism for teaching this process to other prevention personnel does not exist. Anintelligent tutorial expert system is being constructed to provide a means for learning the process and to assist in the creation of specific prevention plans. An intelligent tutoring system (ITS) contains two types of knowledge—domain and tutoring. The domain knowledge for wildfire prevention is structured around several foci: (1) individual concepts used in prevention planning; (2) explicitly specified interrelationships between concepts; (3) deductive methods that contain subjective judgment normally unavailable to less-experienced users; (4) analytical models of fire behavior used for identification of hazard areas; (5) how-to guidance needed for performance of planning tasks; and (6) expository information that provides a rationale for planning steps and ideas. Combining analytical, procedure, inferential, conceptual, and expositional knowledge into a tutoring environment provides the student and/or user with a multiple perspective of the subject matter. Aconcept network provides a unifying framework for structuring and utilizing these diverse forms of prevention planning knowledge. This network structure borrows from and combines semantic networks and frame-based knowledge representations. The flexibility of this organization facilitates an effective synthesis and organization of multiple knowledge forms. |
| |
Keywords: | Artificial intelligence Expert system Knowledge base Tutorial |
本文献已被 SpringerLink 等数据库收录! |
|