This repository contains: - the Content Type Dataset Version 1.5 (in the folder "Datasets"); - the latest version of the guidelines for annotating Content Types; the data statement related to CTD V1.5; - a set of spreadsheets containing metadata about the documents included in the dataset, e.g. year of publication, author's name, author's nationality, author's gender (in the folder "Documents_Metadata"); - the data to replicate a set of experiments for the identification of Content Types (in the folder "Datasets"); - the best model for the identification of Content Types obtained adopting the BiLSTM-CNN-CRF with ELMo-Representations for Sequence Tagging implementation by Nils Reimers and Iryna Gurevych (in the folder "Best_Model"); - the data used to calculate the Inter-Annotator Agreement (in the folder "IAA"): the script used for calculating Cohen's k is available here: https://github.com/johnnymoretti/CAT_R_Kappa_Cohen To replicate the experiments / Run the models To replicate the experiments or run the models, you have to clone and install the BiLSTM-CNN-CRF with ELMo-Representations for Sequence Tagging implementation by Nils Reimers and Iryna Gurevych available at https://github.com/UKPLab/elmo-bilstm-cnn-crf To replicate the experiments, you have to rename the data in Datasets/Cross-Genre and Datasets/Cross-Time to "train.txt", "dev.txt", and "test.txt" To run the model on new data, the input file must be in CoNLL-like format one token per line with empty lines separating sentences. From the BiLSTM-CNN-CRF with ELMo-Representations repository you can use "RunModel_CoNLL_Format.py" Data statement CURATION RATIONALE: We adopt a broad perspective on texts selection assuming that good computational models for NLP must be able to deal with different genres (synchronic dimension) as well as with different times (diachronic dimension). This approach aims at facilitating the re-use of models in different fields of study, and promoting the cross-fertilisation among disciplines, especially in the area of Humanities. On the basis of this approach, we collected texts in English from three different genres: newspaper articles, travel reports, and travel guides. For each of these genres, we collected data published between the second half of the 1800s and the beginning of the 2000s. In designing the corpus, one of our goals was to keep the combination of time and genre as much balanced as possible, in terms of number of tokens and clauses. Furthermore, given the phenomenon under study, we decided to preserve documents' integrity rather than truncating them. LANGUAGE VARIETY: en-GB, en-US, en-AU. ANNOTATOR DEMOGRAPHIC Annotator #1: Age: 36 years old Gender: female Race/ethnicity: caucasian Native language: Italian Socioeconomic status Training in linguistics/other relevant discipline: MA in Computational Linguistics Annotator #2: Age: 37 years old Gender: male Race/ethnicity: caucasian Native language: Italian Socioeconomic status Training in linguistics/other relevant discipline: PhD in Computational Linguistics Annotator #3: Age: 25 years old Gender: female Race/ethnicity: caucasian Native language: Italian Socioeconomic status Training in linguistics/other relevant discipline: MA in Linguistics SPEAKER DEMOGRAPHIC SPEECH SITUATION: Information included in the spreadsheets.
|Date made available||18-May-2021|
|Publisher||University of Groningen|
Keywords on Datasets