An Approach of Similarity Measurement for Heterogeneous and Unordered Mass Customization Products

The heterogeneous and unordered characters of mass customization products cause a huge challenge to analyze the similarity among them. This paper mainly deals with the similarity issues among heterogeneous and unordered mass customization products. Experts’ scoring and local differences are introduced to alleviate the difficulties of evaluating the changes caused by structural operations. And the quantitative correlations between local differences and product level characters such as function or/and performance are established through deep learning. On the basis of these, an approach of similarity analysis for heterogeneous and unordered mass customization products is proposed. It provides practical methods for designer to divide heterogeneous and unordered mass customization products into reasonable groups effectively.