Random Projection Trees Revisited
نویسندگان
چکیده
The Random Projection Tree (RPTREE) structures proposed in [1] are space partitioning data structures that automatically adapt to various notions of intrinsic dimensionality of data. We prove new results for both the RPTREE-MAX and the RPTREE-MEAN data structures. Our result for RPTREE-MAX gives a nearoptimal bound on the number of levels required by this data structure to reduce the size of its cells by a factor s ≥ 2. We also prove a packing lemma for this data structure. Our final result shows that low-dimensional manifolds have bounded Local Covariance Dimension. As a consequence we show that RPTREE-MEAN adapts to manifold dimension as well.
منابع مشابه
Non-crossing trees revisited: cutting down and spanning subtrees
Here we consider two parameters for random non-crossing trees: i the number of random cuts to destroy a sizen non-crossing tree and ii the spanning subtree-size of p randomly chosen nodes in a size-n non-crossing tree. For both quantities, we are able to characterise for n ∞ the limiting distributions. Non-crossing trees are almost conditioned Galton-Watson trees, and it has been already shown,...
متن کاملDifferentially-Private Learning of Low Dimensional Manifolds
In this paper, we study the problem of differentially-private learning of low dimensional manifolds embedded in high dimensional spaces. The problems one faces in learning in high dimensional spaces are compounded in differentially-private learning. We achieve the dual goals of learning the manifold while maintaining the privacy of the dataset by constructing a differentially-private data struc...
متن کاملBertrand’s Paradox Revisited: More Lessons about that Ambiguous Word, Random
The Bertrand paradox question is: “Consider a unit-radius circle for which the length of a side of an inscribed equilateral triangle equals 3 . Determine the probability that the length of a ‘random’ chord of a unit-radius circle has length greater than 3 .” Bertrand derived three different ‘correct’ answers, the correctness depending on interpretation of the word, random. Here we employ geomet...
متن کاملWhich Spatial Partition Trees are Adaptive to Intrinsic Dimension?
Recent theory work has found that a special type of spatial partition tree – called a random projection tree – is adaptive to the intrinsic dimension of the data from which it is built. Here we examine this same question, with a combination of theory and experiments, for a broader class of trees that includes k-d trees, dyadic trees, and PCA trees. Our motivation is to get a feel for (i) the ki...
متن کاملAn algorithm for `1 nearest neighbor search via monotonic embedding
Fast algorithms for nearest neighbor (NN) search have in large part focused on `2 distance. Here we develop an approach for `1 distance that begins with an explicit and exactly distance-preserving embedding of the points into `2. We show how this can efficiently be combined with random-projection based methods for `2 NN search, such as locality-sensitive hashing (LSH) or random projection trees...
متن کامل