Simplistic Collection and Labeling Practices Limit the Utility of Benchmark Datasets for Twitter Bot Detection.
Chris HaysZachary SchutzmanManish RaghavanErin WalkPhilipp ZimmerPublished in: CoRR (2023)
Keyphrases
- benchmark datasets
- uci machine learning repository
- uci repository
- detection method
- automatic detection
- social media
- ensemble methods
- detection accuracy
- detection algorithm
- false alarms
- object detection
- utility function
- detection rate
- case study
- false positives
- prediction accuracy
- document collections
- active learning
- social networks
- micro blogging