Data driven design optimisation: an empirical study of demand discovery combining theory of planned behaviour and Bayesian networks
将贝叶斯网络与计划行为理论结合，用客户调研数据建立模型分析行为意向，指导设计优化，发现增强主观规范和感知行为控制能提高购买或使用概率。
Many theoretical methods have been applied to research user behaviour and requirements. However, the uncertainty associated with customer characteristics often biases the conclusions drawn from customer research and affects the effectiveness of product design. In this paper, Bayesian networks (BN) are introduced into the research on customer behaviour analysis based upon theory of planned behaviour (TPB), and an analysis model driven by customer research data is established from the perspective of user behaviour intention to guide design optimisation. Combining the User background Factor with the TPB Factor, the model analyses the uncertainty of the association between the two, and corrects the errors in the designer's prior knowledge through structural learning. By a case study the paper finds that the evaluations that enhance customers’ subjective norms and perceived behavioural control lead to a greater probability of purchase or use. In addition, customers with specific characteristics are more inclined to generate behaviour intention. The paper finally provides a design optimisation plan based upon the result of the research and discusses about the advantages of the research approaches and the directions of future researches.