改进的有监督的局部线性嵌入算法及实验演示

流形学习方法中的LLE算法可以将高维数据在保持局部邻域结构的条件下降维到低维流形子空间中．并得到与原样本集具有相似局部结构的嵌入向量集合。LLE算法在数据降维处理过程中没有考虑样本的分类信息。针对这些问题进行研究，提出改进的有监督的局部线性嵌人算法（MSLLE），并利用MatLab对该改进算法的实现效果同LLE进行实验演示比较。通过实验演示表明，MSLLE算法较LLE算法可以有利于保持数据点本身内部结构。

The Improved Supervised Locally Linear Embedding Algorithm and Experimental Demonstration

LLE algorithm can remain the nonlinear manifold structure after the task of dimensionality reduction based on manifold learning, and it gets the embedding vector set which similar to local structure of the original sample set. LLE algorithm doesn＇t make use of the classification of information sample point during the process of dimensionality reduction. In view of the disadvantages, proposes a method of modified supervising locally linear embedding. Gives an experimental demonstration compared with demonstrational result of LLE using the MatLab programming. Experimental result shows that MSLLE compared with LLE would conducive to remain the internal structure of the data points themselves.