參考:http://scikit-learn.org/stable/modules/manifold.html
1、流形學習是非線性的降維方法(an approach to non-linear dimensionality reduction)。
2、因為随機映射會随機損失資料内部資訊;因為類似于PCA、LDA等降維方法基于線性假設,經常會損失資料内部非線性的結構資訊;流形學習是線性降維方法的generalization,目的是捕獲資料内部非線性的結構。
3、Though supervised variants exist, the typical manifold learning problem is unsupervised: it learns the high-dimensional structure of the data from the data itself, without the use of predetermined classifications.(盡管是監督的變形,但典型的流形學習還是非監督的)
4、sklearn中對流形學習的實作(了解一下就好了。。。。):
1)Isomap algorithm, short for Isometric Mapping(等距映射)。
2)Locally linear embedding (LLE),局部線性嵌入。 seeks a lower-dimensional projection of the data which preserves distances within local neighborhoods,保持局部線性,但整體非線性,相當于一系列的局部PCA組合。
3)Modified Locally linear embedding
4)Hessian Eigenmapping (also known as Hessian-based LLE: HLLE)
5)Spectral Embedding (also known as Laplacian Eigenmaps)
給個參考:http://blog.csdn.net/chl033/article/details/6107042