中文摘要：

近几年来，流形学习在模式识别、机器学习和数据挖掘等许多领域都受到了广泛的关注．但是，通常的流形学习方法对离群点缺乏鲁棒性对此，提出了一种基于重构权的流形离群点检测方法．该方法在每个样本点构造局部“强”邻域，再利用局部重构权来计算每个样本点的可靠值，最后利用可靠值检测出离群点．该算法具有计算快、参数少、参数敏感性小等优点．基于此离群点检测方法，提出了鲁棒的Isomap算法．实验结果表明，该方法能够有效检测离群点，从而提高流形学习方法对离群点的鲁棒性．国家自然科学基金（10901062）；福建省自然科学科学基金（2010J01336）；华侨大学基本科研业务专项基金

英文摘要：

In past years, the problem with nonlinear dimensionality reduction has aroused a great deal of interest in many research fields, including pattern analysis, machine learning, and data mining. However, the general manifold learning methods are not robust on the outliers. In the paper, an outlier detection method, based on reconstruction weights, is proposed. The proposed algorithm constructs local ＇strong＇ neighborhoods on each sample point, and computes the reliability score of each sample point using local reconstruction weights, and then detects the outliers using the reliability scores. The advantages of the algorithm are that it has fast computation, low parameter, and low parameter sensitivity. Based on the proposed outlier detection method, the robust Isomap algorithm is proposed in this paper. Experimental results illustrate that the proposed algorithm can detect the outliers efficiently and make the manifold learning methods more robust on the outliers.