Record Details

Replication Data for: Learning Supervised Topic Models for Classification and Regression from Crowds

Harvard Dataverse (Africa Rice Center, Bioversity International, CCAFS, CIAT, IFPRI, IRRI and WorldFish)

View Archive Info
 
 
Field Value
 
Title Replication Data for: Learning Supervised Topic Models for Classification and Regression from Crowds
 
Identifier https://doi.org/10.7910/DVN/0EYHTG
 
Creator Rodrigues, Filipe
 
Publisher Harvard Dataverse
 
Description This is the data used in the paper:

Rodrigues, F. and Lourenço, M. and Ribeiro, B. and Pereira, F. C. "Learning Supervised Topic Models for Classification and Regression from Crowds". In IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2017.



It contains supervised learning datasets whose labels were through crowdsourcing platforms, namely Amazon Mechanical Turk, or in some cases, by simulation. These datasets cover various tasks, such as:

- classifying posts and news stories;

- classifying images according to their content;

- predicting number of stars of a given user gave to a restaurant based on the review;

- predicting movie ratings using the text of the reviews.



This data is based on popular benchmark datasets: 20newsgroups, Reuters, LabelMe, we8there, MovieReviews.
 
Subject Computer and Information Science
Crowdsourcing
Amazon Mechanical Turk
Image classification
Text classication
Movie reviews
 
Contributor Rodrigues, Filipe