Home → Magazine Archive → June 2015 (Vol. 58, No. 6) → Evaluation Without Ground Truth in Social Media Research → Abstract

Evaluation Without Ground Truth in Social Media Research

By Reza Zafarani, Huan Liu

Communications of the ACM, Vol. 58 No. 6, Pages 54-60

[article image]

With the rise of social media, user-generated content is available at an unprecedented scale. These massive collections of user-generated content can help researchers better understand the online behavior of billions of individuals. This data also enables novel scientific research in social sciences, anthropology, psychology, and economics at scale.

Scientific research demands reproducible and independently verifiable findings. In social media research, scientific findings can be in the form of behavioral patterns, as in "Individuals commented on this Facebook post due to its quality." To validate these patterns, researchers can survey the individuals exhibiting a pattern to verify if it truly captured their intentions. In data mining terms, such validation is known as "evaluation with ground truth."19 However, social media users are scattered around the world. Without face-to-face access to individuals on social media, is it even possible to perform an evaluation in social media research? That is, how can researchers verify the user behavioral patterns found are indeed the "true patterns" of these individuals?


No entries found