The YouTube Spam Collection v. 1 is a public set of YouTube labeled comments that have been collected for spam research. It has five datasets composed by 1,956 real and non-encoded messages that were tagged as legitimate (ham) or spam.
This corpus has been collected using the YouTube Data API v3.
The samples were extracted from the comments section of 5 videos that were among the 10 most viewed on YouTube during the collection period. The table below lists the 5 datasets collected, the YouTube video ID, the number of samples in each class and the total number of samples per dataset.
| Dataset | YouTube ID | # Spam | # Ham | Total | Link |
|---|---|---|---|---|---|
| Psy | 9bZkp7q19f0 | 175 | 175 | 350 | Link 1 |
| KatyPerry | CevxZvSJLk8 | 175 | 175 | 350 | Link 2 |
| LMFAO | KQ6zr6kCPj8 | 236 | 202 | 438 | Link 3 |
| Eminem | uelHwf8o7_U | 245 | 203 | 448 | Link 4 |
| Shakira | pRpeEdMmmQ0 | 174 | 196 | 370 | Link 5 |
Note: the comments are chronologically sorted.
The collection is composed by one csv file per dataset, where each line has the following attributes:
COMMENT_ID,AUTHOR,DATE,CONTENT,CLASS
We offer one example bellow:
z12oglnpoq3gjh4om04cfdlbgp2uepyytpw0k,Francisco Nora,2013-11-28T19:52:35,please like :D https://premium.easypromosapp.com/voteme/19924/616375350,1
We would appreciate:
- If you find this collection useful, make a reference to the paper below and the web page: http://dcomp.sor.ufscar.br/talmeida/youtubespamcollecion/.
- Send us a message either to talmeida < AT > ufscar.br or tuliocasagrande < AT > acm.org in case you make use of the corpus.
We offer a comprehensive study of this corpus in the following paper. This work presents a number of statistics, studies and baseline results for several machine learning methods.
Alberto, T.C., Lochter J.V., Almeida, T.A. Filtragem Automática de Spam nos Comentários do YouTube. Anais do XII Encontro Nacional de Inteligência Artificial e Computacional (ENIAC'15), Natal, RN, Brazil, 2015. (pending approval)
The YouTube Spam Collection has been created by Tiago A. Almeida, Tulio C. Alberto and Johannes V. Lochter.
© Tiago A. Almeida, Tulio C. Alberto and Johannes V. Lochter, 2015.