By Santosh S. Vempala

Random projection is an easy geometric procedure for decreasing the dimensionality of a suite of issues in Euclidean area whereas retaining pairwise distances nearly. The process performs a key function in different leap forward advancements within the box of algorithms. In different situations, it presents based substitute proofs. The e-book starts off with an ordinary description of the strategy and its simple homes. Then it develops the strategy within the context of purposes, that are divided into 3 teams. the 1st staff involves combinatorial optimization difficulties reminiscent of maxcut, graph coloring, minimal multicut, graph bandwidth and VLSI structure. awarded during this context is the speculation of Euclidean embeddings of graphs. the following team is computer studying difficulties, in particular, studying intersections of halfspaces and studying huge margin hypotheses. The projection technique is extra subtle for the latter software. The final set contains difficulties encouraged through details retrieval, specifically, nearest neighbor seek, geometric clustering and effective low-rank approximation. inspired by means of the 1st functions, an extension of random projection to the hypercube is built the following. during the publication, random projection is used in an effort to comprehend, simplify and fix growth on those vital and likely unrelated difficulties. The ebook is acceptable for graduate scholars and study mathematicians drawn to computational geometry.

Show description

Read or Download The Random Projection Method (Dimacs Series in Discrete Math) PDF

Similar Probability Statistics books

Combinatorics of Experimental Design (Oxford science publications)

This booklet describes direct and recursive equipment for the development of combinatorial designs. it's splendid to the statistician via its dialogue of ways the designs at present utilized in experimental paintings were acquired and during its insurance of alternative identified and probably important designs.

In Defence of Objective Bayesianism

How strongly when you think some of the propositions for you to convey? that's the key query dealing with Bayesian epistemology. Subjective Bayesians carry that it's mostly (though no longer completely) as much as the agent as to which levels of trust to undertake. target Bayesians, nonetheless, keep that acceptable levels of trust are principally (though no longer completely) made up our minds via the agent's facts.

Understanding Data

For information for use through sociologists, and particularly through scholars of sociology, they have to first be effortless to appreciate and use. for that reason this booklet is aimed toward that legion sociologists and scholars who've regularly feared numbers; it employs a lot visible exhibit, for instance, as an ideal way into the knowledge.

Extra info for The Random Projection Method (Dimacs Series in Discrete Math)

Show sample text content

Rated 4.19 of 5 – based on 26 votes