流形学习已成为机器学习和数据挖掘领域的研究热点。比如,算法LLE(Locally Linear Embedding)作为一种非线性降维算法有很好的泛化性能,被广泛地应用于图像分类和目标识别,但其仅仅假设了数据集处于单流形的情况。MM-LLE(Multiple Manif...流形学习已成为机器学习和数据挖掘领域的研究热点。比如,算法LLE(Locally Linear Embedding)作为一种非线性降维算法有很好的泛化性能,被广泛地应用于图像分类和目标识别,但其仅仅假设了数据集处于单流形的情况。MM-LLE(Multiple Manifold Locally Linear Embedding)学习算法作为一种考虑多流形情况的改进算法,依然存在几点不足之处。因此,提出改进的MM-LLE算法,通过任意两类间的局部低维流形组合并构建分类器来提高分类精度;同时改进原算法计算最佳维度的方法。通过与算法ISOMAP、LLE以及MM-LLE比较分类精度,实验结果验证了改进算法的有效性。展开更多
In this paper, we present a more simple and much shorter proof for the main result in the paper ＂ An negative answer to a conjecture on the self-similar sets satisfy- ing the open set condition＂, which was published...In this paper, we present a more simple and much shorter proof for the main result in the paper ＂ An negative answer to a conjecture on the self-similar sets satisfy- ing the open set condition＂, which was published in the journal Analysis in Theory and Applications in 2009.展开更多
文摘流形学习已成为机器学习和数据挖掘领域的研究热点。比如,算法LLE(Locally Linear Embedding)作为一种非线性降维算法有很好的泛化性能,被广泛地应用于图像分类和目标识别,但其仅仅假设了数据集处于单流形的情况。MM-LLE(Multiple Manifold Locally Linear Embedding)学习算法作为一种考虑多流形情况的改进算法,依然存在几点不足之处。因此,提出改进的MM-LLE算法,通过任意两类间的局部低维流形组合并构建分类器来提高分类精度;同时改进原算法计算最佳维度的方法。通过与算法ISOMAP、LLE以及MM-LLE比较分类精度,实验结果验证了改进算法的有效性。
基金Supported in part by National Natural Science Foundation of China (No.10961003)
文摘In this paper, we present a more simple and much shorter proof for the main result in the paper ＂ An negative answer to a conjecture on the self-similar sets satisfy- ing the open set condition＂, which was published in the journal Analysis in Theory and Applications in 2009.
