상세정보
Export to Refworks부가기능
Scalable Algorithms for Mining Dynamic Graphs and Hypergraphs with Applications to Anomaly Detection
상세 프로파일
| 자료유형 | 학위논문 |
|---|---|
| 서명/저자사항 | Scalable Algorithms for Mining Dynamic Graphs and Hypergraphs with Applications to Anomaly Detection. |
| 개인저자 | Ranshous, Stephen Michael. |
| 단체저자명 | North Carolina State University. |
| 발행사항 | [S.l.]: North Carolina State University., 2018. |
| 발행사항 | Ann Arbor: ProQuest Dissertations & Theses, 2018. |
| 형태사항 | 134 p. |
| 기본자료 저록 | Dissertation Abstracts International 79-12B(E). Dissertation Abstract International |
| ISBN | 9780438284029 |
| 학위논문주기 | Thesis (Ph.D.)--North Carolina State University, 2018. |
| 일반주기 |
Source: Dissertation Abstracts International, Volume: 79-12(E), Section: B.
Adviser: Nagiza F. Samatova. |
| 요약 | Graph data mining has become a ubiquitous tool for researchers and practitioners in numerous domains, including social sciences, financial markets, and computer security. In particular, mining dynamic graphs has gained substantial interest in th |
| 요약 | We propose two changes for how anomaly detection is performed over large-scale dynamic graphs to cope with the growing constraints. First, we transition from the traditional approach of analyzing graph streams, where each object in the stream is |
| 요약 | In our first component, based on our extensive survey and gap analysis of the field, we begin with the simplest case, undirected graph edge streams. Key graph properties necessary for our anomaly detection algorithm are approximated from the str |
| 요약 | In our second component, we plan to extend the streaming model from graph edges to hypergraph edges, or hyperedges. As hyperedges represent higher order relationships, not strictly pairwise, we must transition to a more flexible notion of simila |
| 요약 | Finally, our last component examines the potential for pattern based anomalies in dynamic directed hypergraphs (dirhypergraphs). We perform a case study using the Bitcoin network, and propose an edge-based pattern which we posit may represent mo |
| 일반주제명 | Computer science. |
| 언어 | 영어 |
| 바로가기 | 
: 이 자료의 원문은 한국교육학술정보원에서 제공합니다. |
