Multilevel Nonlinear Dimensionality Reduction for Manifold Learning
نویسندگان
چکیده
Nonlinear dimensionality reduction techniques for manifold learning, e.g., Isomap, may become exceedingly expensive to carry out for large data sets. This paper explores a multilevel framework with the goal of reducing the cost of unsupervised manifold learning. In addition to savings in computational time, the proposed multilevel technique essentially preserves the geodesic information, and so it can potentially improve on some manifold learning methods which do not preserve isometry. An application to K-means clustering is also presented. Experimental results indicate that the multilevel approach can be an appealing alternative to standard techniques.
منابع مشابه
مدل ترکیبی تحلیل مؤلفه اصلی احتمالاتی بانظارت در چارچوب کاهش بعد بدون اتلاف برای شناسایی چهره
In this paper, we first proposed the supervised version of probabilistic principal component analysis mixture model. Then, we consider a learning predictive model with projection penalties, as an approach for dimensionality reduction without loss of information for face recognition. In the proposed method, first a local linear underlying manifold of data samples is obtained using the supervised...
متن کاملAlgorithms for manifold learning
Manifold learning is a popular recent approach to nonlinear dimensionality reduction. Algorithms for this task are based on the idea that the dimensionality of many data sets is only artificially high; though each data point consists of perhaps thousands of features, it may be described as a function of only a few underlying parameters. That is, the data points are actually samples from a low-d...
متن کاملNonlinear Manifold Learning 6.454 Summary
Manifold learning is the process of estimating a low-dimensional structure which underlies a collection of high-dimensional data. Here we review two popular methods for nonlinear dimensionality reduction, locally linear embedding (LLE, [1]) and IsoMap [2]. We also discuss their roots in principal component analysis and multidimensional scaling, and provide a brief comparison of the underlying a...
متن کاملNonlinear Dimensionality Reduction
The visual interpretation of data is an essential step to guide any further processing or decision making. Dimensionality reduction (or manifold learning) tools may be used for visualization if the resulting dimension is constrained to be 2 or 3. The field of machine learning has developed numerous nonlinear dimensionality reduction tools in the last decades. However, the diversity of methods r...
متن کاملLarge-Scale Manifold Learning by Semidefinite Facial Reduction
The problem of nonlinear dimensionality reduction is often formulated as a semidefinite programming (SDP) problem. However, only SDP problems of limited size can be directly solved directly using current SDP solvers. To overcome this difficulty, we propose a novel SDP formulation for dimensionality reduction based on semidefinite facial reduction that significantly reduces the number of variabl...
متن کامل