This file is part of a benchmark dataset often cited in studies evaluating bot detection algorithms, such as Botometer (formerly BotOrNot) or similar classifiers [1, 5].
The archive typically contains JSON-formatted metadata for approximately 28 million tweets or a subset of accounts used to train and test machine learning models for identifying automated behavior [4, 6]. twitter028.7z
It is most commonly associated with the following research context: This file is part of a benchmark dataset
It is frequently referenced in the paper "The DARPA Twitter Bot Challenge" or subsequent studies that used the DARPA 2015 dataset to distinguish between human and bot accounts [2, 7]. The filename refers to a specific compressed data
The filename refers to a specific compressed data archive used in several academic research papers focused on Twitter bot detection and social media manipulation [2, 3].
Researchers use this specific file to ensure reproducibility when testing new neural networks or forensic tools against established "gold standard" datasets of known bots [3, 8].