View The Document

Accession Number:

AD1168000

Title:

Mining Large Dynamic Graphs and Tensors

Author(s):

• Shin, Kijung

Author Organization(s):

• Carnegie Mellon University

Report Date:

2019-02-02

Abstract:

Graphs are ubiquitous, representing a variety of information, ranging from who follows whom on online social networks to who reviews what on e-commerce sites. Many of these graphs are large (e.g., online social networks with over two billion active users) and dynamic (i.e., nodes and edges can be added and removed over time). Moreover, they are with rich side information (e.g., e-commerce reviews with timestamps, ratings, and text) and thus naturally modeled as tensors (i.e., multi-dimensional arrays).

Document Type:

Conference:

Journal:

Pages:

321

File Size:

11.59MB

Contracts:

Grants:

Descriptors:

Identifiers:

• graph mining
• tensor mining
• stream mining
• graph stream
• edge stream
• tensor stream
• streaming algorithms
• approximation algorithms
• distributed algorithms
• out-of-core algorithms
• external-memory algorithms
• distributed streaming algorithms
• MapReduce
• Hadoop
• triangle counting
• graph summarization
• graph compression
• tensor decomposition
• tucker decompositionanomaly detection
• fraud detection

SubjectCategory:

• Behavioral and Social Sciences
• Mathematical and Computer Sciences

Communities of Interest:

• Energy and Power Technologies

Distribution Statement:

Approved For Public Release

View The Document