Dynamic Neighborhood Selection for Nonlinear Dimensionality Reduction

Authors:
Yubin Zhan;Jianping Yin;Jun Long
Affiliations:
Computer School, National University of Defense Technology, Changsha, China;Computer School, National University of Defense Technology, Changsha, China;Computer School, National University of Defense Technology, Changsha, China
Venue:
MDAI '09 Proceedings of the 6th International Conference on Modeling Decisions for Artificial Intelligence
Year:
2009

Citing 10
Cited 0

Laplacian Eigenmaps for dimensionality reduction and data representation

Neural Computation
Think globally, fit locally: unsupervised learning of low dimensional manifolds

The Journal of Machine Learning Research
Principal Manifolds and Nonlinear Dimensionality Reduction via Tangent Space Alignment

SIAM Journal on Scientific Computing
Building connected neighborhood graphs for isometric data embedding

Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining
Building k-edge-connected neighborhood graph for distance-based data projection

Pattern Recognition Letters
Building k-Connected Neighborhood Graphs for Isometric Data Embedding

IEEE Transactions on Pattern Analysis and Machine Intelligence
Selection of the optimal parameter value for the Isomap algorithm

Pattern Recognition Letters
Building Connected Neighborhood Graphs for Locally Linear Embedding

ICPR '06 Proceedings of the 18th International Conference on Pattern Recognition - Volume 04
Riemannian Manifold Learning

IEEE Transactions on Pattern Analysis and Machine Intelligence
A more topologically stable locally linear embedding algorithm based on R*-tree

PAKDD'08 Proceedings of the 12th Pacific-Asia conference on Advances in knowledge discovery and data mining

Quantified Score

Hi-index	0.00

Visualization

Abstract

Neighborhood construction is a necessary and important step in nonlinear dimensionality reduction algorithm. In this paper, we first summarize the two principles for neighborhood construction via analyzing existing nonlinear dimensionality reduction algorithms: 1) data points in the same neighborhood should approximately lie on a low dimensional linear subspace; and 2) each neighborhood should be as large as possible. Then a dynamic neighborhood selection algorithm based on this two principles is proposed in this paper. The proposed method exploits PCA technique to measure the linearity of a finite points set. Moreover, for isometric embedding, we present an improved method of constructing neighborhood graph, which can improve the accuracy of geodesic distance estimation. Experiments on both synthetic data sets and real data sets show that our method can construct neighborhood according to local curvature of data manifold and then improve the performance of most manifold algorithms, such as ISOMAP and LLE.