An agent-based model is developed to understand the behaviors and rule-sets that generate social media networks. Simple rules are used to generate a network that resembles a backcloth (friend/follow) network collected at an earlier date using the Twitter API. Model parameter adjustments were made to reproduce the collected network’s summary stat…
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore
CCDF-Fit- Collected.png
CCDF-Fit- Generated.png
CCDF-Fit- collected-comparison.png
CCDF-Fit- generated-comparison.png
Distribution Comparison - KS Distance.xlsx
KS-Ditance - x min - Collected.png
KS-Ditance - x min - Generated.png
LICENSE
Network_Fitting.ipynb
README.md

README.md

fitting-a-powerlaw

An agent-based model is developed to understand the behaviors and rule-sets that generate social media networks. Simple rules are used to generate a network that resembles a backcloth (friend/follow) network collected at an earlier date us-ing the Twitter API. Model parameter adjustments were made to reproduce the collected network’s summary statistics and stylized specifics such as average de-gree, clustering, community size and distribution, as well as general structural met-rics. An approximate network was produced in line with the general properties of our collected data. In this paper, we focus on and discuss the properties of the de-gree distribution of our generated network and compare using recently developed methods for analyzing power law and heavy tailed distributions. It is proposed that a power law distribution forms a more appropriate fit for both our collected and generated networks then an exponential distribution, that a power law distri-bution is a more appropriate fit for smaller, more clustered networks than larger networks with multiple active local social forces. Finally, that seldom used models of fit such as a lognormal or a truncated power law distribution outperform both the exponential and power law distributions for either network.