CONTEXT RELIABILITY AND INTERNAL VALIDITY OF AN EXPERT-SYSTEM TO ASSESS LANDSCAPE VISUAL VALUES

Citation
Gj. Buhyoff et Pa. Miller, CONTEXT RELIABILITY AND INTERNAL VALIDITY OF AN EXPERT-SYSTEM TO ASSESS LANDSCAPE VISUAL VALUES, AI applications, 12(1-3), 1998, pp. 76-82
Citations number
10
Categorie Soggetti
Computer Science Artificial Intelligence","Environmental Sciences","Computer Science Artificial Intelligence",Forestry,Agriculture
Journal title
ISSN journal
10518266
Volume
12
Issue
1-3
Year of publication
1998
Pages
76 - 82
Database
ISI
SICI code
1051-8266(1998)12:1-3<76:CRAIVO>2.0.ZU;2-O
Abstract
An expert system, designed to assess the visual quality of natural lan dscape scenes, was subjected to set of validation and context reliabil ity tests. This system was evaluated for its context reliability by te sting its performance when it was used on-site to evaluate natural lan dscapes versus its performance off-site using photographs of the same sites. System internal validity was evaluated by comparing the results of the on- and off-site evaluations against the professional judgment of an expert landscape architect of the visual quality of these lands capes. The data indicate that the Explanation of Visual Assessment (EV A) expert system provides a means of evaluating landscape quality with high context reliability and internal validity.