类标约束下的非负矩阵分解算法研究-电子与通信工程专业论文.docxVIP

AbstractIII Abstract III Abstract Nonnegative matrix factorization (NMF), with its psychological intuition of combining parts to form a whole, has become a popular method in finding parts-based, linear representation of non-negative data. It attracts many researchers’ attention and has been widely used in data mining, computer vision and pattern recognition problems. In this thesis, we emphasize the application of NMF in label constrained problem. For clustering or classification problems, auxiliary constraints are often used in NMF in order to enforce desired characteristics to the computed s