Bagging and random feature subsets are used simultaneously to achieve maximum diversity among k-NN component experts in a fusion system. Experimental results indicate that for certain data the combination of the two design methods in a fusion system is beneficial. We also compare two random feature subset design methods, a widely used, conventional, expert based method and a system based design...

This paper provides the first solution to the kinetic reverse k-nearest neighbor (RkNN) problem in R, which is defined as follows: Given a set P of n moving points in arbitrary but fixed dimension d, an integer k, and a query point q / ∈ P at any time t, report all the points p ∈ P for which q is one of the k-nearest neighbors of p.

3.1 Vector M odel ............................................................................................. 4 3.2 Terms W eighting ....................................................................................... 4 3.3 Similarity M easuring ................................................................................. 5 4 Improving kNN Performance using Different Term Weighting Sch...

3D human action recognition is an important current challenge at the heart of many research areas lying to the modeling of the spatio-temporal information. In this paper, we propose representing human actions using spatio-temporal motion trajectories. In the proposed approach, each trajectory consists of one motion channel corresponding to the evolution of the 3D position of all joint coordinat...

Several types of nearest neighbor (NN) search have been proposed and studied in the context of spatial databases. The most common type is the point NN query, which retrieves the nearest neighbors of an input point. Such a query, however, is usually meaningless in highly dynamic environments where the query point or the database objects move/change over time. In this paper we study alternative f...

Nonlinear dimensionality reduction methods often rely on the nearest-neighbors graph to extract low-dimensional embeddings that reliably capture the underlying structure of high-dimensional data. Research however has shown that computing nearest neighbors of a point from a highdimensional data set generally requires time proportional to the size of the data set itself, rendering the computation...

We provide finite-sample analysis of a general framework for using k-nearest neighbor statistics to estimate functionals of a nonparametric continuous probability density, including entropies and divergences. Rather than plugging a consistent density estimate (which requires k → ∞ as the sample size n → ∞) into the functional of interest, the estimators we consider fix k and perform a bias corr...

The performance of a fuzzy k-NN rule depends on the number k and a fuzzy membership-array W[I, mR], where l and m R denote the number of classes and the number of elements in the reference set X R respectively. The proposed learning procedure consists in iterative finding such k and W which minimize the error rate estimated by the 'leaving one out' method.

Assume we are given a sample of points from some underlying distribution which contains several distinct clusters. Our goal is to construct a neighborhood graph on the sample points such that clusters are “identified”: that is, the subgraph induced by points from the same cluster is connected, while subgraphs corresponding to different clusters are not connected to each other. We derive bounds ...