자료유형  

 학위논문

Control Number  

0017164126

International Standard Book Number  

9798384455752

Dewey Decimal Classification Number  

020

Main Entry-Personal Name  

Tandon, Aditya.

Publication, Distribution, etc. (Imprint  

[S.l.] : Indiana University., 2024

Publication, Distribution, etc. (Imprint  

Ann Arbor : ProQuest Dissertations & Theses, 2024

Physical Description  

207 p.

General Note  

Source: Dissertations Abstracts International, Volume: 86-03, Section: B.

General Note  

Advisor: Fortunato, Santo.

Dissertation Note  

Thesis (Ph.D.)--Indiana University, 2024.

Summary, Etc.  

요약This dissertation explores the development and application of network science techniques to complex systems, focusing on two primary areas: community detection in networks and the evolution of science.In community detection, two key challenges are addressed. First, it tackles the scalability limitations of consensus clustering methods by proposing a fast consensus clustering algorithm. This approach leverages sparse sampling techniques and local network structure, achieving significant computational savings while maintaining stability and accuracy. The algorithm enables the analysis of large-scale networks previously intractable. The thesis also explores the potential of graph embedding techniques, which have gained recent popularity, for community detection. Through a comprehensive comparative analysis, various embedding methods are evaluated against state-of-the-art community detection algorithms on benchmark graphs. The findings reveal that certain embedding techniques, when optimally tuned, can perform comparable to or slightly better than traditional algorithms. Challenges related to parameter selection and computational complexity are also identified and discussed.The dynamics of scientific progress is studied via the analysis of networks at different levels. At the micro level, an investigation of the author citation network reveals that the simple preferential attachment model does not capture the bursty dynamics of citation accumulation. Instead, a mechanistic model based on recency is proposed, which successfully reproduces the distributions of both citation counts and temporal bursts. At the macro level, the evolution of science is examined by studying the dynamics of a knowledge space, obtained by embedding papers in geometric space through language models. The analysis reveals that papers are based on more diverse concepts now than in the first half of the past century, indicating that scholars increasingly use tools and techniques from other disciplines to solve their problems. This demonstrates that broad, cross-disciplinary information search has become a defining feature of global science, with multidisciplinary journals playing a crucial role in broadcasting ideas across the scientific system.These contributions advance both methodological approaches in network science and our understanding of dynamics in science. The developed techniques have potential applications in large-scale network analysis across domains, while the insights into scientific evolution carry implications for research evaluation and science policy.

Subject Added Entry-Topical Term  

Information science.

Subject Added Entry-Topical Term  

Systems science.

Subject Added Entry-Topical Term  

Engineering.

Subject Added Entry-Topical Term  

Computer science.

Subject Added Entry-Topical Term  

Information technology.

Index Term-Uncontrolled  

Citation dynamics

Index Term-Uncontrolled  

Community detection

Index Term-Uncontrolled  

Complex networks

Index Term-Uncontrolled  

Graph embeddings

Index Term-Uncontrolled  

Science of science

Added Entry-Corporate Name  

Indiana University Informatics

Host Item Entry  

Dissertations Abstracts International. 86-03B.

Electronic Location and Access  

로그인을 한후 보실 수 있는 자료입니다.

Control Number  

joongbu:658614