Abstract: To solve the matter of the integration and selection of multidimensional demands in innovative design of products, a method based on three-dimensional coordinates and Cosine similarity measure is proposed. First of all, three-dimensional coordinates is used to show vector of innovation demands. Secondly, the intensity of every innovation demand is calculated by vector module, and the similarity between innovation demands and benchmark innovation demand is calculated by Cosine similarity measure, based on the intensity and similarity, a quadrifid graph model can be built and it can make an intuitive selection. Depending on the innovative integration degree we can make a precise selection. After combing the quadrifid graph model and innovative integration degree, a systematic selecting method is obtained. Finally, through the example, it proves the feasibility of the method.

Key words: three-dimensional coordinates of innovation demand, benchmark demand, Cosine similarity, innovative integration degree

摘要： 为解决产品创新设计过程中多维需求的集成与筛选问题,现提出一种基于三维坐标及余弦相似度的创新需求筛选方法。首先采用创新需求三维坐标来表示每项创新需求,其次应用矢量模的计算方法确定每项创新需求强度的大小,根据余弦相似度计算创新需求与基准创新需求间的相似度,以需求强度和相似度为基础,构建一个创新需求四分图来对创新需求进行直观的筛选,计算出每项创新需求的创新集成度来对创新需求做出精确的筛选,将二者相结合组成系统的筛选方法,最后通过实例验证了所提方法的有效性。

关键词: 创新需求三维坐标, 基准创新需求, 余弦相似度, 创新集成度