KEYWORD GENERATION SYSTEM BASED ON DESIGN MODEL AND CO-OCCURRENCE WORD LEARNING

抄録

This study focuses on one of the designers' thinking models: Multispace design model and developed a design support tool based on the reasoning defined in the model. The system utilizes the Word2vec (language modeling and feature learning techniques in natural language processing) and Interpretive Structural Modeling for generating the linguistic stimuli for idea generation (design keywords) and automatically developing a design keyword structure, respectively. The proposed system was used by the eight students in the design experiment to confirm its applicability. The results show that the number of the generated design keywords and ideas in the conditions supported by the proposed system is increased. Additinoally, the questionnaire about the design cognition during the tasks reveals the proposed system support the design activity. These suggest the applicability of the proposed system as a design support tool.