流形学习的主要目标是发现嵌入在高维数据空间的低维光滑流形1 近年来基于谱图理论的学

习算法受到研究者的广泛关注1 介绍了流形与流形学习的关系,着重研究了几种有代表性的基于谱图

理论的流形学习算法,并对算法进行了比较分析,最后进行总结和对进一步的研究做了展望1-The main objective of manifold learning is to find embedded in high-dimensional data space of a low-dimensional smooth manifold in recent years, a learning algorithm based on spectral theory of attention by the researchers introduced a manifold and manifold learning relationship, focused on the several representative theoretical spectra based manifold learning algorithms, and a comparative analysis algorithm, and finally a summary and further research done on the prospect 1