Patterns and Surprises in Rich but Noisy Network Data

Explore the intricacies of network analysis in this seminar by Mark Newman, focusing on social and biological networks. Delve into the challenges of estimating network structure from rich but noisy data, including measurement errors, contradictory observations, and missing information. Examine how the pattern of errors in network data can provide valuable insights into both the data itself and the underlying systems. Learn about expectation maximization techniques, clustering coefficients, and their applications in various network types such as social networks, food webs, and plant-pollinator networks. Discover how to evaluate network analysis methods using ground truth, recall, and precision metrics. Gain a deeper understanding of network structure estimation and its implications for research in fields ranging from internet studies to biological systems.

Introduction
Social networks
Ordinary data
True structure
Data
Expectation maximization
Expectations maximization
Network structure
More powerful objects
Example
The catch
Example application
Ground truth
Recall and precision
Net result
Clustering coefficient
Food web data
Experiments
Plant pollinator network
EM algorithm
Friendship network