Graduation Year

2023

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Computer Science and Engineering

Major Professor

Les A. Piegl, Ph.D.

Committee Member

Kandethody Ramachandran, Ph.D.

Committee Member

Richard Segall, Ph.D.

Committee Member

Robert Karam, Ph.D.

Committee Member

Tempestt Neal, Ph.D.

Keywords

Attention Mechanism, Deep Learning, Entity Detection, Graph Convolution Network

Abstract

With the development of transportation network, social network, and communication network, there are many applications in streaming data. For example, traffic congestion happens between the origin and destination of daily trips. Traffic analysis can help plan the trips so that traffic congestion can be avoided. Social network and communication network represent the behaviors of the entire population. People build connections based on their hobbies, daily activities, photos, videos, simple messages, and even anonymous web surfing. All of these can be turned into commercial use, such as product marketing, business network building, and technology trending. Data science is about how to model data for data issues, domain specific patterns, etc. If the sample set is big enough and the data is relevant, it is possible to engineer this process and to generate results. Once the data model is built, we can fit the model with the data and run proper algorithms to get answers. However, the challenges can be from data store, sample quality to information extraction. Especially for graph analysis, it needs to deal with not only the valuation of the vertices but also the connections between vertices, and how many connections each vertex can have. According to empirical experiments, the distributions of vertices, edges, and derived measurements, such as closeness, betweenness, and clustering coefficient have different distributions. When we work on data modeling strategies, both graph properties and topology types need to be under. In this dissertation, we discuss how to efficiently perform information extraction from graphs. Graphs can be considered as the structural representation of the social networks regarding the natural properties of information, such as recency, relevance, valuation, and validation. We focus on data loading, graph sampling, and application development.

Download

Share

COinS