Predicting the Internet’s Evolution with Decision Trees and Lasso Logistic Regression Model

Description

This paper creates a model of the internet as a network graph and then extracts its structural characteristics; these characteristics in turn help us build models that can predict how the Internet will evolve. Lasso logistic regression, gradient-boosted decision trees and random forest decision trees were evaluated in their robustness in predicting the internet's evolution. For details, see a short paper here

 

Undergraduate Research Project conducted by Cuong Nguyen `14

 

Data and Source Code

Data for this project were collected by The Cooperative Association for Internet Data Analysis.

Source code can be found here