Granularity Enhancement of Extracted Preferential Probabilities From Design Team Discussion

Abstract

Preferences are a formal way to represent a designer's choices when assigning priorities for a set of possible design alternatives within the context of the design process. A design team's preferences can change over the life of project, and knowledge of this evolution can be useful for understanding a team's rationale as well as its confidence in a decision. This paper presents a "sliding window" approach (SPPT) to the extraction of preference related information from transcribed design team discussion. The approach suggested in this paper can assess design preferences over time with a finer granularity than a previous approach known as PPT, and removes perturbations that occur when there is little design team discussion. Both SPPT and PPT were applied to a discussion transcript. Results show good consistency among SPPT, PPT and survey results. SPPT is also able to detect more changes in design team preference.