How Am I Doing 2: Computing the Language of Appraisal in Design

There has been a wide body of design research about the categorisation of design text by the content or role of the text. However, recent research on information use in design has suggested the need to consider the semantic orientation toward the subject matter because the orientation might signal design rationale and other critical meta-functions. This paper presents a computational implementation of a natural language classifier for the understanding of appraisal in design text, that is, the expression of an opinion or attitude toward the designed work, the design process, and the design team. Using the supervised machine learning technique of support vector machines and data from informal and formal design documents, including transcripts of conversations between designers, we find that machine learning techniques can be trained to categorise the semantic meaning of design documents and the orientation of the opinion expressed in the document.