Abstract: As a novel dimensionality reduction technique，manifold learning algorithm aims at discovering the lower dimensional manifold structure which is embedded in the high dimensionality data space，and derives an efficient lower dimensional representation.At present，manifold learning has become a hot issue in the field of pattern recognition，machine learning and data mining.The fundamental ideas，some latest research resutls and algorithm analysis of manifold learing are introduced.Besides，some problems for further research are presented and analyzed.

Key words: manifold learning, dimensionality reduction, pattern recognition, machine learning

摘要： 流形学习算法作为一种新的维数降维方法工具，其目标是发现嵌入在高维数据空间中的低维流形结构，并给出一个有效的低维表示。目前，流形学习已成为模式识别、机器学习和数据挖掘领域的研究热点问题。介绍了流形学习的基本思想、一些最新研究成果及其算法分析，并提出和分析了有待进一步研究的问题。

关键词: 流形学习, 维数降维, 模式识别, 机器学习