基于SVD协同过滤的同风格图片推荐算法

Style consistence is the most important thing in UI-design process. In order to reduce the cost of maintaining style consistence of UI for designers, this paper proposes a recommendation algorithm of pictures having the same style based on SVD collaborative filtering. This algorithm computes rating matrix by 0 or 1 according to whether the images are used or not, and it calculates the similarity of pictures which have been used and which have not during each new design process to predict whether they are in the same style, and optimize the calculation by SVD. It is shown in the experimental results that this algorithm is effective and it can recommend pictures of high-quality after being optimized by SVD.