Representation of structured data of the text genre as a technique for automatic text processing

Fonseca, Claudia Aparecida; Guelpeli, Marcus Vinícius Carvalho; Souza Netto, Rafael Santiago de

doi:10.35699/1983-3652.2022.35445

home Sumário navigate_before Anterior Atual Seguinte navigate_next

home Sumário

Articles • Texto livre 15 • 2022 • https://doi.org/10.35699/1983-3652.2022.35445 linkcopy

Representation of structured data of the text genre as a technique for automatic text processing

Authorship SCIMAGO INSTITUTIONS RANKINGS

Abstract

The present article was developed in the field of Natural Language Processing and Language Studies based on a corpus compiled by computational tools. This study is based on the assumption that it is helpful to trace a close relationship between corpus generation/annotation and the assessment of the constitutive elements of the text genre source. It aims to demonstrate, through specific studies of structured data from the text genre ‘scientific article’, alternatives to automatic text processing techniques. In order to reach the intended goal, the authors created a computational model for the compilation of a linguistic, specialized Corpus, representative of the genre Scientific Article CorpACE. The object of study includes the constitutive elements of scientific articles, marked in XML, extracted and collected from the SciELO-Scientific Electronic Library Online database. The final product was a database obtained with information extracted and structured in XML format, which designates and identifies the markups of the genre being analyzed and is available for many tools and applications. The results demonstrate how the representation of constitutive elements of the genre can condense available information with hierarchical and dynamic processes built during the compilation. At the end of the study, it is believed that more research will be required for bringing Language Science and Computer Science closer with emphasis on NLP in the attempt to represent and manipulate linguistic knowledge in its many levels - morphological, syntactic, semantic and discursive - in order to improve implementation and manipulation of automatic text processing.

Keywords:
Corpus linguistics; Natural language processing; Scientific article; Text genre; Corpora annotation

Technique	Approach	Description	Publications
Removal of stopwords - (filter)	Linguistic	Filtering process for removing words with little relevance, in the attempt to measure all the information that does not constitute knowledge in the text. The idea behind this filter is to remove words that contain little or no content, such as articles, prepositions, pronouns, conjunctions, adverbs, numerals and interjections. Additionally, terms that commonly or rarely appear are probably not significantly relevant and thus can also be removed.	Luhn (1958), Salton and McGill (1983), Frakes and Baeza-Yates (1992), Lui, Li, and Choy (2007) and De Oliveira Júnior and Esmin (2012).
TF-IDF - (Term Frequency - Inverse Document Frequency)	Statistical	Term Frequency (TF) is based on the premise that the importance of a term appearing in a document is directly proportional to its occurrence. Inverse Document Frequency (IDF) is based on the premise that the specificity of a term can be measured by an inverse function of the number of documents in which it occurs. Therefore, this technique consists in weighing the importance of each term within a background corpus, normally consisting of documents belonging to the same domain and then eliminating the list of very common words.	Luhn (1958), Jones (1972), Bhatia and Jaiswal (2015), Liu, Li, and Feng (2017) and Rocha and Guelpeli (2017).
Latent Semantic Analysis (LSA)	Hybrid	A method that uses synonyms and polysemy to extract and represent the semantic meaning of words in a context. This representation is obtained from mathematical calculations and applications that analyze the relationship between terms and documents, decomposing them into index vectors.	Landauer, Foltz, and Laham (1998) and Scarton and Aluísio (2010).
N-grams	Statistical	This technique consists of word co-occurrence and permits a statistical prediction of two or more terms in a text appearing in a certain sequence. An n-gram is a contiguous substring of n items of a given sequence of text or speech.	Cohen (1995), Liu, Webster, and Kit (2009), L. F de Alencar (2010), A. F. de Alencar (2013a) and Tonelli and Pianta (2011).
Segmentation	Hybrid	Segmenting the content of the text in individualized sentences, which represents a minimal semantic set for defining a proposition.	Lin, Hsieh, and Chuang (2009), SOUSA, KEPLER, and FARIA (2010) and A. F. de Alencar (2013b).
Tokenization (Text segmentation in words)	Hybrid	Consists in the process that segments a sequence of text characters into a sequence of significant units (words) that compose the text. The spaces and punctuation are generally adopted as delimiting tokens for western languages.	Webster and Kit (1992), SOUSA, KEPLER, and FARIA (2010), A. F. de Alencar (2013b) and Silva, Trindade, et al. (2015)
Lemmatization and stemming	Linguistic	Lemmatization is the representation of each word of the input text in its primitive form (lemma). The process of word stemming has the purpose of removing suffixes and prefixes from a term, so that it is reduced to its radical (stem).	Lovins (1968), SOUSA, KEPLER, and FARIA (2010) and Rolim, Ferreira, and Costa (2016).
Part-of-Speech (POS) Tagging	Linguistic	It consists of tagging the words of the input text with their respective grammatical classes and syntactic distributions. The main morphosyntactic tagging techniques include: Rule-based, which follows tagging rules manually coded by linguists; Probabilistics, which employs statistical tagging methods in which each word has a finite set of possible tags and is labeled with the most probable tag; and Hybrid, which involves the combination of rule-based and probabilistic techniques.	Lau et al. (2008), Domingues, Favero, and De Medeiros (2008), SOUSA, KEPLER, and FARIA (2010), A. F. de Alencar (2013b) and Santos and Zadrozny (2014).
Text genre Tagging	Linguistic	It consists in tagging the main characteristics of the input text genre. This technique permits building the structural model in an arboreal format and adding linguistic data; information about the relationships between elements of the production context, or sentences or sentence fragments of the general text infrastructure; and the visualization of the constitutive dimensions of the base genre. This tagging can delimit various constitutive elements of the text genre, such as: bibliographic references, sections, abstract, paragraphs, tables, figures, funding, title, subtitles, authorship, keywords, and many others. This technique can retrieve the basic structure of the input text by planning the root nodes and their possible affiliations, which represent the textual infrastructure. The preprocessing of a genre will somehow be influenced by the recognition of the superstructure and infrastructure of its compositional organization.	Fonseca (2018).

Value	Description
Cases	Cases/case studies
Conclusions	Conclusions/final considerations/final remarks
Discussion	Discussions
Introduction	Introduction/synopsis
Materials	Materials used
Methods	Methodology /methods
Results	Results

Statistics	Total
Number of articles	87
Total word count	750380
Word count of smallest article	6144
Word count of largest article	18758
Average word count per article	8625
Word count filtered from constitutive elements of genre	130128
Word count of Corpus EdReal	383506
Word count of Corpus EdPesq	366874

vertical_align_top show_chart

more_horiz close
- image
- translate
- link
- article
- vertical_align_top
- file_download
- show_chart
- image
- translate
- link
- article

location_on

Universidade Federal de Minas Gerais - UFMG Av. Antônio Carlos, 6627 - Pampulha, Cep: 31270-901, Belo Horizonte - Minas Gerais / Brasil, Tel: +55 (31) 3409-6009 - Belo Horizonte - MG - Brazil
E-mail: revistatextolivre@letras.ufmg.br

rss_feed Acompanhe os números deste periódico no seu leitor de RSS

[TFN4] Source: Created by the author.