Publications
Detailed Information
A method getting the transitive kNN graph of a kNN graph preserving topology
DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 강명주 | - |
dc.contributor.author | 박정호 | - |
dc.date.accessioned | 2020-10-13T04:02:21Z | - |
dc.date.available | 2020-10-13T04:02:21Z | - |
dc.date.issued | 2020 | - |
dc.identifier.other | 000000162183 | - |
dc.identifier.uri | https://hdl.handle.net/10371/170703 | - |
dc.identifier.uri | http://dcollection.snu.ac.kr/common/orgView/000000162183 | ko_KR |
dc.description | 학위논문 (석사) -- 서울대학교 대학원 : 자연과학대학 수리과학부, 2020. 8. 강명주. | - |
dc.description.abstract | In this paper, we recall some previous research of a bijection between transitive
directed graphs and topologies on a finite set. In data science area, k Nearest Neighbor(kNN) graph is a widely used graph. It can be interpreted by a directed graph in the sense of in and out degrees. kNN graph is not transitive in general. Adding or eliminating edges, there are various methods to get a transitive directed graph from a given kNN graph. In this paper, we prove that there is a unique way which make a transitive directed graph preserving topology. Moreover, we suggest a simple method to make the transitive directed graph without computing the topology of a given kNN graph. | - |
dc.description.abstract | 이 학위 논문에서는 유한집합 위에서의 위상들과 transitive 유향그래프의 일대일대응에 관련된 내용들을 상기한다. kNN 그래프는 데이터를 활용한 과학 분야에서 널리 사용된다. 우리는 kNN그래프의 in-degree와 out-degree를 기준으로 kNN그래프를 유향그래프로 해석할 수 있다. 하지만 일반적으로 kNN그래프는 transitive 유향그래프가 아니고 주어진 kNN그래프의 간선을 추가하거나 빼는 등 다양한 방법으로 transitive 유향그래프를 만들 수 있다. 이 학위 논문을 통해서 우리는 위상을 보존하는 transitive 유향그래프를 얻는 방법이 유일함을 증명하고 구체적인 방법을 제시한다. 더 나아가서 위상의 계산 없이 transitive kNN그래프를 구하는 계산 알고리즘을 제안한다. | - |
dc.description.tableofcontents | 1 Introduction, 1
2 Preliminaries, 2 2.1 Notions of finite directed graphs, 2 2.2 Adjacency matrices, 4 2.3 Notions of finite topological spaces, 6 2.4 A Bijection between transitive directed graph and topology, 8 2.5 Local Maximum Feature selection Algorithm, 10 3 Algorithm getting a transitive directed graph from a directed graph, 12 3.1 A transitive directed subgraph preserving corresponding topology, 12 3.2 Explicit rule to compute the transitive directed subgraph preserving corresponding topology, 14 3.3 Computation algorithm via adjacency matrix,15 4 Transitive kNN subgraph of a kNN graph, 17 5 Conclusion, 18 References, 19 Abstract (in Korean), 20 | - |
dc.language.iso | eng | - |
dc.publisher | 서울대학교 대학원 | - |
dc.subject | kNN graph | - |
dc.subject | transitive directed graph | - |
dc.subject | finite topology | - |
dc.subject | kNN그래프 | - |
dc.subject | transitive 유향그래프 | - |
dc.subject | 유한위상 | - |
dc.subject.ddc | 510 | - |
dc.title | A method getting the transitive kNN graph of a kNN graph preserving topology | - |
dc.type | Thesis | - |
dc.type | Dissertation | - |
dc.contributor.department | 자연과학대학 수리과학부 | - |
dc.description.degree | Master | - |
dc.date.awarded | 2020-08 | - |
dc.identifier.uci | I804:11032-000000162183 | - |
dc.identifier.holdings | 000000000043▲000000000048▲000000162183▲ | - |
- Appears in Collections:
- Files in This Item:
Item View & Download Count
Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.