Vis enkel innførsel

dc.contributor.advisorHetland, Magnus Lie
dc.contributor.authorPhilipp, Andre
dc.date.accessioned2019-09-11T10:55:59Z
dc.date.created2015-06-15
dc.date.issued2015
dc.identifierntnudaim:13144
dc.identifier.urihttp://hdl.handle.net/11250/2615815
dc.description.abstractWe present one and develop two heuristic algorithms for multi-focal clustering. The first algorithm is a local search algorithm based on the Partition Around Medoids (PAM) algorithm. The second algorithm is a modification of the sparse spatial selection (SSS) to add multi-focality. The third algorithm uses single link cluster analysis (SLCA) to form the clusters. Further we develop a Mixed Integer Program to find the optimal solution. We show that these algorithms have different applicable properties that can be utilized for different data sets and with different preconditions. We compare the heuristics and the multi-focal solutions to existing heuristics and uni-focal, ball-regions, and show that the multi-focal regions can be advantageous to reduce the number of comparisons by using a Monte Carlo algorithm to simulate searching in the indexes.en
dc.languageeng
dc.publisherNTNU
dc.subjectDatateknologi, Komplekse datasystemeren
dc.titleClustering with Multi-Focal Clustersen
dc.typeMaster thesisen
dc.source.pagenumber74
dc.contributor.departmentNorges teknisk-naturvitenskapelige universitet, Fakultet for informasjonsteknologi og elektroteknikk,Institutt for datateknologi og informatikknb_NO
dc.date.embargoenddate10000-01-01


Tilhørende fil(er)

Thumbnail
Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel