Dataset: A Cross-Platform and Cross-Interaction Study of User Personality based on Images on Twitter and Flickr
Version 2 2018-06-09, 18:41Version 2 2018-06-09, 18:41
Version 1 2018-06-09, 18:40Version 1 2018-06-09, 18:40
dataset
posted on 2018-06-09, 18:41authored byZahra Riahi Samani, Sharath GuntukuSharath Guntuku, Mohsen Ebrahimi Moghaddam, Daniel Preoțiuc-Pietro, Lyle Ungar
This folder contains the following dataset: Crossed-Linked Flickr and Twitter dataset.
Psycho-Flickr consists of a set of users who answered the BFI survey. We collected profile and up to 300 posted and liked pictures for each user. Please contact Christina Segalin (http://www.cristinasegalin.com/) for access to these labels.
Crossed-Linked Flickr and Twitter consists of a set of users with active accounts both on Flickr and Twitter.
We used text mining approaches to predict personality traits for this set of users.
We collected profile and up to 300 posted and liked picture for each user.
Twitter or Flickr user ids with their text-predicted/BFI survey Big-Five personality scores are presented.
Features extracted from profile images and averaged over posted and Liked Images are presented that include :
Colors Features
CNN Generic Features: 4096 dim penultimate layer features of VGG_Net
CNN object and scene categories:VGG_Net prediction on 1000 objects and 365 scene categories