On The Correlation Clustering Problem: Algorithm for Correlation Clustering Problem

Penumatcha, Sriram

ISBN 10: 3838313542 / ISBN 13: 9783838313542

Published by LAP LAMBERT Academic Publishing, 2009

About this Item

LIKE NEW. SHIPS FROM MULTIPLE LOCATIONS.

Seller Inventory # ERICA80038383135426

Bibliographic Details

• Title: On The Correlation Clustering Problem: Algorithm for Correlation Clustering Problem
• Publisher: LAP LAMBERT Academic Publishing
• Publication Date: 2009
• Language: English
• ISBN 10: 3838313542
• ISBN 13: 9783838313542
• Binding: paperback
• Condition: Like New
• Book Type: book

About this title

Synopsis

We consider the correlation clustering problem which was initially introduced by Bansal, Blum, Chawla et al. Given a complete graph G on n vertices, with weights of +1 or -1 defined on the edges, we want to find a partition which maximizes the sum of the number of edges with positive weights inside the clusters plus the number of edges with negative weights between different clusters. In this thesis we present a deterministic polynomial time approximation scheme for finding such a partition. Our approach is different from the one given by Bansal, Blum, Chawla et al. as it relies on the Szemeredi's Regularity Lemma. We start by introducing the problem, then we introduce the concepts of regularity lemma and give a proof of Szemeredi's Regularity Lemma. Then we present the algorithm and the proof of the correctness of the algorithm.