一种面向图像分类的流形学习降维算法

A MANIFOLD LEARNING DIMENSIONALITY REDUCTION ALGORITHM FOR IMAGE CLASSIFICATION

数据挖掘中的流形学习降维算法可以应用于图像分类等领域。提出一种面向图像分类的流形学习降维算法Mod-LLE(Modified Locally Linear Embedding)。该算法是针对高维数据的局部线性嵌入降维算法的改进,其整合了图像识别信息来更好地改善优化效果,达到在处理过程中保证原始数据固有的拓扑组成结构。以标准数据集作为案例进行测试。图像分类功能测试与降维性能测试结果表明:该算法对于人脸图像的分类精度比较高,降维性能良好。

Manifold learning dimensionality reduction algorithm is a important tools for data mining such as image classfication applications.This paper proposed a manifold learning dimensionality reduction algorithm called Mod-LLE(Modified Locally Linear Embedding)for image classification.It is a improved modify version of locally linear embedding algorithm.It integrated the identification information of the data to keep the correlation composition of initiate data and explored the intrinsic topology to hunting hidden relationships of the data.A seriers of experiments were done for testing Mod-LLE using standard data sets.The experiments results show that the algorithm can obtain a high classification accuracy for image classification,and it has a better dimensionality reduction performance.