train_sentences: Sentence corpus - train part
In fastrtext: 'fastText' Wrapper for Text Classification and Word Representation

Description Usage Format Details Source

This corpus contains sentences from the abstract and introduction of 30 scientific articles that have been annotated (i.e. labeled or tagged) according to a modified version of the Argumentative Zones annotation scheme.

1	train_sentences

2 data frame with 3117 rows and 2 variables:

text: the sentences as a character vector
class.text: the category of the sentence

These 30 scientific articles come from three different domains:

PLoS Computational Biology (PLOS)
The machine learning repository on arXiv (ARXIV)
The psychology journal Judgment and Decision Making (JDM)

There are 10 articles from each domain. In addition to the labeled data, this corpus also contains a corresponding set of unlabeled articles. These unlabeled articles also come from PLOS, ARXIV, and JDM. There are 300 unlabeled articles from each domain (again, only the sentences from the abstract and introduction). These unlabeled articles can be used for unsupervised or semi-supervised approaches to sentence classification which rely on a small set of labeled data and a larger set of unlabeled data.

===== References =====

S. Teufel and M. Moens. Summarizing scientific articles: experiments with relevance and rhetorical status. Computational Linguistics, 28(4):409-445, 2002.

S. Teufel. Argumentative zoning: information extraction from scientific text. PhD thesis, School of Informatics, University of Edinburgh, 1999.

https://archive.ics.uci.edu/ml/index.php

fastrtext documentation built on Oct. 30, 2019, 11:32 a.m.