Font Size: a A A

Analysis Of Hierarchical Architecture Based On Distance

Posted on:2015-02-27Degree:MasterType:Thesis
Country:ChinaCandidate:L T PengFull Text:PDF
GTID:2268330425474422Subject:Applied Mathematics
Abstract/Summary:PDF Full Text Request
In this paper, normalization distances are focused on as main objects based on thegranular space theory, and the clustering structural analysis is proposed and discussed. Thispaper is organized as follows:In chapter one, the developments of data mining, granular compute, cluster analysis andnormalization distances are briefly outlined both at home and abroad. Besides, the main workand innovations of this paper are given.In chapter two, the representation of the granular space and an improving algorithm tocompute the clustering structure are firstly given. Secondly, the minimum spanning tree isbuilt to explain the generation process of the granular space. At the same time, the concept ofkey point point is introduced and other concepts are defined. Finally, the distance betweeneach point is provided.In chapter three, the clustering structural analysis on the breeding distribution area ofRed-crowned Crane based on5climate factors in Northeast China from1981to1990. Firstly,the characteristics analysis on the related climate data is studied by using the theory andmethods of interval statistics, hierarchical clustering based on normalization distance andclustering fusion technology. Then, the results shows that the most appropriate time inbreeding habitat of Red-crowned Crane. The conclusion coincides with the biologicalcharacteristics and ecological characteristics. This set of research methods possess generalityand will provide basis for the research about the impact of future climate change on thepredictive distribution of the migratory species in Northeast China.In chapter four, at first, data of climate scenarios20C3M(1951-2000), future climatescenarios SRESA1B (2001-2100) and the work in chapter3are extracted by using the theoryand methods of hierarchical clustering based on normalization distance and clustering fusiontechnology. Then, by using MATLAB to design and run program under ArGIS9.3, thepredicting tables about the optimal adaptable residence time, the medium adaptable one andthe general adaptable one of Red-crowned Crane in Northeast China from2041to2050andfrom2091to2100are obtained. By analyzing these tables, the breeding habitats of theRed-crowned Crane will be no longer suitable for Red-crowned Crane breeding under climatechange, furthermore,3factors influencing the residence time of Red-crowned Crane on theirbreeding habitats are related to temperature.In chapter five, the whole work is summarized and reported. Besides, the ideas andthoughts of further research are pointed out.
Keywords/Search Tags:granular space, cluster analysis, normalization distance, hierarchicalclustering, optimal cluster, clustering fusion
PDF Full Text Request
Related items