Keywords

Diffusion of Information, Influence, Clustering Coefficient, Agent-Based Modeling, Transfer Entropy, Network Models

Abstract

Information diffusion is significant in fields such as propagation prediction and influence maximization, with applications in viral marketing and rumor control. Despite conceptual differences, existing diffusion models may not represent identical underlying generative structures. A classification of diffusion of information models is developed based on infection requirements and stochasticity. The study involves analyzing seven existing DOI models on directed scale-free networks. The distinctive properties of each model are identified through simulations and analysis of experimental results. Our analysis reveals that similarity in conceptual design does not imply similarity in behavior concerning speed, the final state of nodes and edges, and sensitivity to parameters. Therefore, we highlight the importance of considering the unique behavioral characteristics of each model when selecting a suitable information diffusion model for a particular application. We further investigate how the network structure and clustering affect the diffusion of information. Our findings reveal that clustering does not consistently accelerate the spread of information. Instead, the extent to which clustering facilitates the dissemination of information is influenced by the interplay between the specific network structure types and the information diffusion model employed. Another significant aspect of information diffusion is the effect of influential nodes. Identifying highly influential nodes is of great interest for strategic targeting in various applications such as viral marketing and information campaigns. Our follow-up study aims to identify influential nodes using a transfer entropy-based method. In this work, we use our method to identify influential users in Twitter data and compare the results against other existing methods. Finally, we developed a methodology based on Transfer Entropy to evaluate influence in the context of information diffusion. This methodology demonstrated its superiority in predicting user adoption against retweet-based metrics, marking it as a direct and reliable metric for understanding influential users and information diffusion trends.

Completion Date

2024

Semester

Spring

Committee Chair

Garibay, Ivan

Degree

Doctor of Philosophy (Ph.D.)

College

College of Engineering and Computer Science

Department

Industrial Engineering and Management Systems

Degree Program

Industrial Engineering

Format

application/pdf

Identifier

DP0028288

URL

https://purls.library.ucf.edu/go/DP0028288

Language

English

Rights

In copyright

Release Date

May 2024

Length of Campus-only Access

None

Access Status

Doctoral Dissertation (Open Access)

Campus Location

Orlando (Main) Campus

Accessibility Status

Meets minimum standards for ETDs/HUTs

Download

Share

COinS