市场驱动的产品设计中，客户需求往往来自于客户调研。然而，有价值的客户需求也蕴含于海量的在线评论中。但海量的在线评论往往以文本形式存在，这就使得在线评论通常很少直接用于产品设计中。此项研究讨论了基于质量功能展开进行需求分析的一个重要的问题：如何确定产品工程特征的重要性。本研究利用在线评论中丰富的情感信息，提出了一种有序分类的方法。此方法将客户情感偏好作为特征，并将客户整体满意度作为目标，并利用结对比较的方式最终确定产品工程特征的权重。此外，本研究还提出了一个整数线型规划模型。该模型用于将结对比较的结果转化为原始的客户满意度。最终，本研究以亚马逊网站中四个品牌的打印机评论为例，并利用两名产品设计人员的分析结果，进行了一系列探索性实例调研。这些实验证明了提出本研究提出的方法的优点。

In market-driven product design, customer requirements (CRs) are usually obtained from consumer surveys. However, valuable CRs can also be found in a large number of online reviews. Largely due to their free text nature and the quantity, these reviews are often neglected and are seldom utilised directly by designers. In this work, one important question in quality function deployment on how to prioritise engineering characteristics (ECs) is investigated. Customer opinions concerning ECs are extracted from online reviews. By taking advantage of such opinion information, an ordinal classification approach is proposed to prioritise ECs. In a pairwise manner, in which customer opinions are deemed as features and the overall customer satisfactions are regarded as the target values, the weights of ECs are derived. Furthermore, an integer linear programming model is implemented to convert the pairwise results into the original customer satisfaction ratings. Finally, an exploratory case study is presented using reviews of four branded printers collected from Amazon and their analysis was conducted by two experienced design engineers. The experimental study reveals the merits of the proposed approach.

Jian Jin, Ping Ji and Ying Liu. Prioritising engineering characteristics based on customer online reviews for quality function deployment [J]. Journal of Engineering Design. Vol. 25, Issue 7-9, pages 303-324, 2014