Predicting Window View Preferences Using the Environmental Information Criteria

Daylighting standards provide an assessment method that can be used to evaluate the quality of window views. As part this evaluation process, designers must achieve five environmental information criteria (location, time, weather, nature, and people) obtain excellent view. To best our knowledge, these have not yet been verified their scientific validity remains conjectural. In a two-stage experiment, total 451 persons evaluated six view images. Using machine learning models, we found could accurate predictions for preferences. When one was largely preferred over other, accuracy decision tree models ranged from 83% 90%. For smaller differences in preference, 67%. ratings given increased, so did evaluations psychological restoration positive affect. Although causation established, role most important predicting preferences, with nature generally outweighed others. We recommend use practice, but suggest some alterations emphasize importance within design. Instead only supporting high-quality views, should promoted across all thresholds dictating quality.