Shijian Luo

A study of product family design DNA based on product style

In order to solve the problem of product family appearance customization in the product family platform, an approach of product family design DNA based on product style was proposed. The concept and related studies of product family design DNA were discussed, and the relationship between product style and product family design DNA was analyzed. With the study of the primary method of constructing product family design DNA with style factors, a concept model was put forward. Take the lock product family design as an example, a configuration model and computer-aided lock design system were set up, including three-dimensional form design component, perceptual image evaluation component and knowledge base component. Based on the system, the lock design can be accomplished rapidly and the product style-based product family design DNA was test.

Style matching model-based recommend system for online shopping

In order to enrich the recommendation of e-commerce functionality and optimize the user experience of online shopping, a new style matching model-based recommend method of e-commerce was proposed. The Content-Based Image Retrieval(CBIR) technology was used to extract the image feature of color, texture and shape of product, calculated the commodity cluster, which have the similar feature of image content, then was applied to the style matching model, recommend to the user in accordance with the users cognitive style. The style matching mode-based recommendation engine was developed, and integrated into the e-commerce recommend system to help users complete the cognitive of style and guide the online shopping. Finally, a shoe buying recommend system was developed to verify the theory.

A fast collaborative filtering algorithm for implicit binary data

Item-based and user-based collaborative filtering are two well-known algorithms for recommender system in E-commerce. Both the algorithms make use of similarity matrix whose elements represent the similarity of each item pairs or user pairs. A fast algorithm for item-based similarity matrix computation using cosine similarity metric was reviewed and applied for user-based one with some modification. The results show that the fast algorithm can blend well with other similarity metrics, and it can greatly improve the computational performance.