Suchergebnisse

In March 2020, the World Health Organization announced the COVID-19 outbreak as a pandemic. Most previous social media related research has been on English tweets and COVID-19. In this study, we collect approximately 1 million Arabic tweets from the Twitter streaming API related to COVID-19. Focussing on outcomes that we believe will be useful for Public Health Organizations, we analyse them in three different ways: identifying the topics discussed during the period, detecting rumours, and predicting the source of the tweets. We use the k-means algorithm for the first goal with k=5. The topics discussed can be grouped as follows: COVID-19 statistics, prayers for God, COVID-19 locations, advise and education for prevention, and advertising. We sample 2000 tweets and label them manually for false information, correct information, and unrelated. Then, we apply three different machine learning algorithms, Logistic Regression, Support Vector Classification, and Naïve Bayes with two sets of features, word frequency approach and word embeddings. We find that Machine Learning classifiers are able to correctly identify the rumour related tweets with 84% accuracy. We also try to predict the source of the rumour related tweets depending on our previous model which is about classifying tweets into five categories: academic, media, government, health professional, and public. Around (60%) of the rumour related tweets are classified as written by health professionals and academics.

Creating, curating and maintaining modern political corpora is becoming an ever more involved task. As interest from various socialbodies and the general public in political discourse grows so too does the need to enrich such datasets with metadata and linguisticannotations. Beyond this, such corpora must be easy to browse and search for linguists, social scientists, digital humanists and thegeneral public. We present our efforts to compile a linguistically annotated and semantically tagged version of the Hansard corpus from1803 right up to the present day. This involves combining multiple sources of documents and transcripts. We describe our toolchainfor tagging; using several existing tools that provide tokenisation, part-of-speech tagging and semantic annotations. We also provide anoverview of our bespoke web-based search interface built on LexiDB. In conclusion, we examine the completed corpus by looking atfour case studies making use of semantic categories made available by our toolchain.

Large scale financial narrative processing for UK annual reports has only become possible in the last few years with our prior work on automatically understanding and extracting the structure of unstructured PDF glossy reports. This has levelled the playing field somewhat relative to US research where annual reports (10-K Forms) have a rigid structure imposed on them by legislation and are submitted in plain text format. The structure extraction is just the first step in a pipeline of analyses to examine disclosure quality and change over time relative to financial results. In this paper, we describe and evaluate the use of similar Information Extraction and Natural Language Processing methods for extraction and analysis of annual financial reports in a second language (Portuguese) in order to evaluate the applicability of our techniques in another national context (Portugal). Extraction accuracy varies between languages with English exceeding 95%. To further examine the robustness of our techniques, we apply the extraction methods on a comprehensive sample of annual reports published by UK and Portuguese non-financial firms between 2003 and 2015.

This study combines quantitative semi-automated corpus methods with manual qualitative analysis to investigate the use of Violence metaphors for cancer and end of life in a 1,500,000-word corpus of data from three stakeholder groups in healthcare: patients, family carers and healthcare professionals. Violence metaphors in general, especially military metaphors, are conventionally used to talk about illness, particularly cancer. However, they have also been criticized for their potentially negative implications. The use of innovative methodology enables us to undertake a more rigorous and systematic investigation of Violence metaphors than has previously been possible. Our findings show that patients, carers and professionals use a much wider set of Violence-related metaphors than noted in previous studies, and that metaphor use varies between interview and online forum genres and amongst different stakeholder groups. Our study has implications for the computer-assisted study of metaphor, metaphor theory and analysis more generally, and communication in healthcare settings.

This study combines quantitative semi-automated corpus methods with manual qualitative analysis to investigate the use of Violence metaphors for cancer and end of life in a 1,500,000-word corpus of data from three stakeholder groups in healthcare: patients, family carers and healthcare professionals. Violence metaphors in general, especially military metaphors, are conventionally used to talk about illness, particularly cancer. However, they have also been criticized for their potentially negative implications. The use of innovative methodology enables us to undertake a more rigorous and systematic investigation of Violence metaphors than has previously been possible. Our findings show that patients, carers and professionals use a much wider set of Violence-related metaphors than noted in previous studies, and that metaphor use varies between interview and online forum genres and amongst different stakeholder groups. Our study has implications for the computer-assisted study of metaphor, metaphor theory and analysis more generally, and communication in healthcare settings.

ParlaMint 2.1 is a multilingual set of 17 comparable corpora containing parliamentary debates mostly starting in 2015 and extending to mid-2020, with each corpus being about 20 million words in size. The sessions in the corpora are marked as belonging to the COVID-19 period (after November 1st 2019), or being "reference" (before that date). The corpora have extensive metadata, including aspects of the parliament; the speakers (name, gender, MP status, party affiliation, party coalition/opposition); are structured into time-stamped terms, sessions and meetings; with speeches being marked by the speaker and their role (e.g. chair, regular speaker). The speeches also contain marked-up transcriber comments, such as gaps in the transcription, interruptions, applause, etc. Note that some corpora have further information, e.g. the year of birth of the speakers, links to their Wikipedia articles, their membership in various committees, etc. The corpora are encoded according to the Parla-CLARIN TEI recommendation (https://clarin-eric.github.io/parla-clarin/), but have been validated against the compatible, but much stricter ParlaMint schemas. This entry contains the ParlaMint TEI-encoded corpora with the derived plain text version of the corpus along with TSV metadata on the speeches. Also included is the 2.0 release of the data and scripts available at the GitHub repository of the ParlaMint project. Note that there also exists the linguistically marked-up version of the corpus, which is available at http://hdl.handle.net/11356/1431.

ParlaMint 2.1 is a multilingual set of 17 comparable corpora containing parliamentary debates mostly starting in 2015 and extending to mid-2020, with each corpus being about 20 million words in size. The sessions in the corpora are marked as belonging to the COVID-19 period (from November 1st 2019), or being "reference" (before that date). The corpora have extensive metadata, including aspects of the parliament; the speakers (name, gender, MP status, party affiliation, party coalition/opposition); are structured into time-stamped terms, sessions and meetings; with speeches being marked by the speaker and their role (e.g. chair, regular speaker). The speeches also contain marked-up transcriber comments, such as gaps in the transcription, interruptions, applause, etc. Note that some corpora have further information, e.g. the year of birth of the speakers, links to their Wikipedia articles, their membership in various committees, etc. The corpora are encoded according to the Parla-CLARIN TEI recommendation (https://clarin-eric.github.io/parla-clarin/), but have been validated against the compatible, but much stricter ParlaMint schemas. This entry contains the linguistically marked-up version of the corpus, while the text version is available at http://hdl.handle.net/11356/1432. The ParlaMint.ana linguistic annotation includes tokenization, sentence segmentation, lemmatisation, Universal Dependencies part-of-speech, morphological features, and syntactic dependencies, and the 4-class CoNLL-2003 named entities. Some corpora also have further linguistic annotations, such as PoS tagging or named entities according to language-specific schemes, with their corpus TEI headers giving further details on the annotation vocabularies and tools. The compressed files include the ParlaMint.ana XML TEI-encoded linguistically annotated corpus; the derived corpus in CoNLL-U with TSV speech metadata; and the vertical files (with registry file), suitable for use with CQP-based concordancers, such as CWB, noSketch Engine or KonText. Also included is the 2.1 release of the data and scripts available at the GitHub repository of the ParlaMint project. As opposed to the previous version 2.0, this version corrects some errors in various corpora and adds the information on upper / lower house for bicameral parliaments. The vertical files have also been changed to make them easier to use in the concordancers.