News Corpus Research Articles

A corpus-assisted approach to discursive news values analysis

The main aim of this paper is the elaboration of an analytical tool for comparative studies. For this purpose, I used a combination of Discursive News Values Analysis (DNVA) and Corpus Linguistics (CL) to analyse a corpus of British Broadsheets’ news coverage of the Brexit campaign. The four major British broadsheets which were analysed were The Guardian, The Independent, The Times, and The Daily Telegraph. A specific procedure was designed following previous studies on the topic and considering the challenges and opportunities that such a mixed-method approach (DNVA and CL) can face in exploring journalistic discursive practices and mapping the cultural and ideological discourses around certain topics. Some initial results of the case study are presented to show how the suggested procedure works in practice. From the present study’s findings, the procedure seems to work in a reliable manner, although some challenges should be considered and addressed in future studies.

Developing a synthetic news corpus to validate generic frame detection methods

Frames are a central concept in communication research. Based on our literature review, we propose that frame identification is an act of identifying selected reality and communicative intention. We then highlight the conceptual and methodological issues of frame identification using computational methods. To avoid the correlation between topics and frames, we provide a synthetic dataset for evaluating frames found in multi-topical news content, using the detection of generic frames as a test case. With this dataset, for the first time, we benchmark manual coding and various automatic and semi-supervised methods. Based on the preliminary benchmark results, this study provides evidence that generic frame identification using both manual coding and automatic methods might not be accurate.

Afghan Women’s Representation in News Discourse during the First 365 Days of Taliban Governance A Corpus-Based Analysis

The study attempts to investigate the 365 days of Afghan women in the post-Taliban regime in Afghanistan in news discourse. The study particularly aims to analyze the lexical collocation of the keywords related to Afghan women in the news discourse. The study is qualitative as it analyses the data qualitatively. Corpus Linguistics has been used as a methodology in this study. Fifty online news articles have been selected from Pakistani newspaper The News International, and a corpus has been developed with the name Afghan Women News Corpus. The data have been analysed using LancsBox v.5.x (Brezina et al., 2020). The study draws on Baker’s (2004) Corpus-Assisted Discourse Analysis framework. The corpus was uploaded to LancsBox and a list of collocations of the keyword Afghan Women was generated. Both the left and right collocates of the keywords were analyzed in detail and were used for thematic development and coding. The findings of this analysis revealed the basic patterns the data used to deal with the topic of Afghan women in particular and their social setup in general. It is found that in the majority of cases, the news discourse negatively represents the Taliban regime, occasionally shows it neutrally, and rarely shows it positively. The research concludes with a discussion of the main findings by highlighting the discourse surrounding community representation before and after the war.

Redundant Acronym Syndrome in Indonesian News Articles: A Corpus Analysis Approach

The aims of the study are to explore the occurrence of Redundant Acronym Syndrome (RAS) in Indonesian news articles and examine its implications by employing corpus analysis approach. The main source of the data from the Indonesian news corpus, which consists of 150,466 articles from Indonesian news websites over a six-month period spanned from July 2015 to December 2015. The data analysis was conducted by using Antconc software version 4.02. The analysis reveals a systematic and recognisable structure in acronym formation, highlighting cultural or organisational conventions such as initial phoneme retention, retention of syllables + letters, and retention of syllables and syllables, and retention of initial phonemes and letters. The findings of this study revealed a total of 40 Redundant Acronym with 5730 occurrences identified in the Indonesian news corpus, covering various fields such as the name of football clubs, banks, political parties, educational institutions, government-related terms, and commonly used phrases. This research also discussed that that RAS in news articles can have both negative and positive impacts. It negatively impact readability by making repetitions and potentially confusing the reader. In contrast, RAS can also improve clarity by emphasising certain aspects or reinforcing associations in acronyms, so that readers can instantly recognise and understand them. The findings provided valuable insights for news writers and editors, emphasising the importance of a balance between readability and clarity in news articles.

Forecasting CPI with multisource data: The value of media and internet information

AbstractUsing a large Chinese news corpus and Internet search data, we conduct an in‐depth out‐of‐sample forecasting study of Consumer Price Index (CPI) with the monthly macroeconomic database. For this purpose, we combine penalized regression and mixed‐frequency data sampling (MIDAS) methods to deal with the mixed‐frequency and high‐dimensional problems. Then we measure the time‐varying contributions of data from different sources and industries via dynamic model averaging. The results show that different types of data all provide enhancement in CPI forecasting, and the inclusion of media and Internet data further improves CPI forecasts' accuracy and timeliness due to their thorough coverage and fast update. In particular, the importance of alternative data is highlighted when the economy experiencing downturn or high uncertainty.

A Comparison of Chinese-English News Metaphors for Corpus Analysis in the Context of Big Data Technology

Abstract In this paper, firstly, the Chinese-English news corpus is constructed using crawler technology, and a distributed retrieval system is designed to search the Chinese-English data in the corpus. Secondly, the potential semantic analysis algorithm is improved to extract semantic information from the Chinese-English corpus and realize the semantic search data source. Finally, by designing the experimental environment as well as the samples of the Chinese-English news corpus, the performance of the corpus as well as the distribution retrieval performance are analyzed, and the comparison of Chinese-English news metaphors is explored. The results show that the system’s word-level accuracy is 96.32% when the training corpus size reaches 100,000 word-level, the F-value is 96.08, and the latency of 1Mb/s is higher than the latency of 3Mb/s and the latency time of 5M/s, and the corresponding latency is always stabilized within 1.0. The proportion of conceptual metaphors in Chinese news is above 0.5, while in English news, the proportion is mostly around 0.3. Most of the Chinese news is conceptual metaphors, while most English news is interpersonal metaphors.

From Star Trek to The Hunger Games: emblem gestures in science fiction and their uptake in popular culture

Abstract Research on emblems to date has not drawn on corpus methods that use public data. In this paper, we use corpus methods to explore the use of original fictional gestures in the real world. We look at two examples from popular science fiction, the Vulcan salute from Star Trek and the three-finger salute from The Hunger Games. First, a Twitter corpus of the Vulcan salute emoji shows that it is used to represent Star Trek fandom and wider nerd culture, alongside its use as a greeting. Second, a global news corpus shows the three-finger salute has come to be used as a pro-democracy protest gesture across political and cultural boundaries in South East Asia. These corpus studies show different trajectories for the two gestures, with the three-finger salute escaping the confines of its fictional world, while the Vulcan salute has come to stand in as a reference to the media it originated from. We conclude with a reflection on the opportunities, challenges, and limitations of bringing corpus methods to gesture studies.

Alignment and antagonism in flux: A diachronic sentiment analysis of attitudes towards the Chinese mainland in the Hong Kong press

Despite extensive discussions about Hong Kong’s sentiments towards the Chinese mainland since the sovereignty transfer in 1997, there has been a lack of large-scale empirical analysis of attitude change in the mainstream media, which both reflect and shape sentiments in society. To address this gap, the present study employs an optimised semantic-based automatic sentiment analysis method to examine a corpus of news about China from 1997 to 2020 in three leading Chinese-language newspapers in Hong Kong, namely Apple Daily, Ming Pao, and Oriental Daily News. The analysis reveals that although the Hong Kong press had a positive emotional tone towards China in general, the overall sentiment was becoming increasingly negative. Meanwhile, the alignment and antagonism towards China have both increased, providing empirical evidence for attitudinal polarisation in the Hong Kong press. Specifically, Apple Daily’s depictions of China became increasingly negative, though with some positive turns before 2008, whereas Oriental Daily News consistently expressed favourable sentiments. Ming Pao maintained an impartial stance towards China through an increased but balanced representation of positive and negative sentiments, with its subjectivity and sentiment intensity growing to an industry-standard level. By explicating the cross-newspaper and cross-period variations, the results provide new insights into the complexity of sentiments towards China in the Hong Kong press, and media attitudes in general in terms of the “us” and “them” positionings.

Tracking diachronic sentiment change of economic terms in times of crisis: Connotative fluctuations of 'inflation' in the news discourse.

The present study focuses on the fluctuation of sentiment in economic terminology to observe semantic changes in restricted diachrony. Our study examines the evolution of the target term 'inflation' in the business section of quality news and the impact of the Great Recession. This is carried out through the application of quantitative and qualitative methods: Sentiment Analysis, Usage Fluctuation Analysis, Corpus Linguistics, and Discourse Analysis. From the diachronic Great Recession News Corpus that covers the 2007-2015 period, we extracted sentences containing the term 'inflation'. Several facts are evidenced: (i) terms become event words given the increase in their frequency of use due to the unfolding of relevant crisis events, and (ii) there are statistically significant culturally motivated changes in the form of emergent collocations with sentiment-laden words with a lower level of domain-specificity.

Online Satirical News as Instances of Liquid Racism: Evidence from Greek

When discussing the relationship between humour and racism, research usually concentrates on ethnic or racist jokes or on other humorous texts which expressly target migrants or other minorities. Relevant studies more or less explicitly ascribe to critical humour studies investigating topics such as how and why humour targeting the linguistic, cultural, or religious Other reproduces and maintains social discrimination and inequality, as well as how and why the generic conventions of humorous genres do not incite the audience to think critically of their content but instead enhance their tolerance for discriminatory and racist standpoints. In this context, the present study explores humorous texts that at first sight appear to have antiracist intentions. More specifically, I analyse a corpus of satirical news coming from popular Greek websites and targeting majority people for their racist views and practices towards migrants. The analysis is based on the concepts of script opposition and target as defined in the framework of the General Theory of Verbal Humour. The findings suggest that distinguishing between antiracist and racist interpretations is not an easy or straightforward matter: humour seems to blur the boundary between racism and antiracism. In order to account for this dimension of humour, I exploit the concept of liquid racism put forward by Weaver (The Rhetoric of Racist Humour) to account for the ambiguities of humorous discourse when it involves racist and antiracist meanings and interpretations.

DAT-MT Accelerated Graph Fusion Dependency Parsing Model for Small Samples in Professional Fields.

The rapid development of information technology has made the amount of information in massive texts far exceed human intuitive cognition, and dependency parsing can effectively deal with information overload. In the background of domain specialization, the migration and application of syntactic treebanks and the speed improvement in syntactic analysis models become the key to the efficiency of syntactic analysis. To realize domain migration of syntactic tree library and improve the speed of text parsing, this paper proposes a novel approach-the Double-Array Trie and Multi-threading (DAT-MT) accelerated graph fusion dependency parsing model. It effectively combines the specialized syntactic features from small-scale professional field corpus with the generalized syntactic features from large-scale news corpus, which improves the accuracy of syntactic relation recognition. Aiming at the problem of high space and time complexity brought by the graph fusion model, the DAT-MT method is proposed. It realizes the rapid mapping of massive Chinese character features to the model's prior parameters and the parallel processing of calculation, thereby improving the parsing speed. The experimental results show that the unlabeled attachment score (UAS) and the labeled attachment score (LAS) of the model are improved by 13.34% and 14.82% compared with the model with only the professional field corpus and improved by 3.14% and 3.40% compared with the model only with news corpus; both indicators are better than DDParser and LTP 4 methods based on deep learning. Additionally, the method in this paper achieves a speedup of about 3.7 times compared to the method with a red-black tree index and a single thread. Efficient and accurate syntactic analysis methods will benefit the real-time processing of massive texts in professional fields, such as multi-dimensional semantic correlation, professional feature extraction, and domain knowledge graph construction.

Fake news detection using machine learning: an adversarial collaboration approach

PurposePurveyors of fake news perpetuate information that can harm society, including businesses. Social media's reach quickly amplifies distortions of fake news. Research has not yet fully explored the mechanisms of such adversarial behavior or the adversarial techniques of machine learning that might be deployed to detect fake news. Debiasing techniques are also explored to combat against the generation of fake news using adversarial data. The purpose of this paper is to present the challenges and opportunities in fake news detection.Design/methodology/approachFirst, this paper provides an overview of adversarial behaviors and current machine learning techniques. Next, it describes the use of long short-term memory (LSTM) to identify fake news in a corpus of articles. Finally, it presents the novel adversarial behavior approach to protect targeted business datasets from attacks.FindingsThis research highlights the need for a corpus of fake news that can be used to evaluate classification methods. Adversarial debiasing using IBM's Artificial Intelligence Fairness 360 (AIF360) toolkit can improve the disparate impact of unfavorable characteristics of a dataset. Debiasing also demonstrates significant potential to reduce fake news generation based on the inherent bias in the data. These findings provide avenues for further research on adversarial collaboration and robust information systems.Originality/valueAdversarial debiasing of datasets demonstrates that by reducing bias related to protected attributes, such as sex, race and age, businesses can reduce the potential of exploitation to generate fake news through adversarial data.

Lightweight Scheme to Capture Stock Market Sentiment on Social Media Using Sparse Attention Mechanism: A Case Study on Twitter

Over through the years, people have invested in stock markets in order to maximize their profit from the money they possess. Financial sentiment analysis is an important topic in stock market businesses since it helps investors to understand the overall sentiment towards a company and the stock market, which helps them make better investment decisions. Recent studies show that stock sentiment has strong correlations with the stock market, and we can effectively monitor public sentiment towards the stock market by leveraging social media data. Consequently, it is crucial to develop a model capable of reliably and quickly capturing the sentiment of the stock market. In this paper, we propose a novel and effective sequence-to-sequence transformer model, optimized using a sparse attention mechanism, for financial sentiment analysis. This approach enables investors to understand the overall sentiment towards a company and the stock market, thereby aiding in better investment decisions. Our model is trained on a corpus of financial news items to predict sentiment scores for financial companies. When benchmarked against other models like CNN, LSTM, and BERT, our model is “lightweight” and achieves a competitive latency of 10.3 ms and a reduced computational complexity of 3.2 GFLOPS—which is faster than BERT’s 12.5 ms while maintaining higher computational complexity. This research has the potential to significantly inform decision making in the financial sector.

CORPUS-ASSISTED DISCOURSE STUDY (CADS): WOMEN WITH DISABILITIES IN KOMPAS.COM’S NEWS REPORTING

The interaction of gender and disability has gained attention in academia and activism, emphasizing the importance of combating prejudice and stigma encountered by women with disabilities. The portrayal of this group in the media is critical because it has the potential to either perpetuate stereotypes or contribute to constructive societal change. The study, "Corpus-assisted discourse study (CADS): Women with disabilities in Kompas.com's News Reporting," addressed how one of Indonesia's prominent news websites, Kompas.com, portrayed and discussed women with disabilities in its news reporting. This study conducted a descriptive qualitative study concerning the corpus-aided discourse study (CADS) approach. It used the web scraping of Octoparse to build a sizable news corpus from Kompas.com. Then, AntcConc, a corpus toolkit, generated 26,398 tokens from news items with the keyword phrase perempuan disabilitas (women with disabilities). The findings indicated 45 occurrences of perempuan (women) and disabilitas (disability) associated with the lexical words: ganda (multiple), seksual (sexual), kekerasan (violence), diperkosa (raped), etc. Meanwhile, concordance analysis of the corpus revealed different topic clusters in the news reports: violent crimes and challenges relating to policies for women with disabilities. This study also shed insights on the unintentional potential ramifications for readers' cognitive biases in the discourse that Kompas.com propagated through news reporting.

Adapting Feature Selection Algorithms for the Classification of Chinese Texts

Text classification has been highlighted as the key process to organize online texts for better communication in the Digital Media Age. Text classification establishes classification rules based on text features, so the accuracy of feature selection is the basis of text classification. Facing fast-increasing Chinese electronic documents in the digital environment, scholars have accumulated quite a few algorithms for the feature selection for the automatic classification of Chinese texts in recent years. However, discussion about how to adapt existing feature selection algorithms for various types of Chinese texts is still inadequate. To address this, this study proposes three improved feature selection algorithms and tests their performance on different types of Chinese texts. These include an enhanced CHI square with mutual information (MI) algorithm, which simultaneously introduces word frequency and term adjustment (CHMI); a term frequency–CHI square (TF–CHI) algorithm, which enhances weight calculation; and a term frequency–inverse document frequency (TF–IDF) algorithm enhanced with the extreme gradient boosting (XGBoost) algorithm, which improves the algorithm’s ability of word filtering (TF–XGBoost). This study randomly chooses 3000 texts from six different categories of the Sogou news corpus to obtain the confusion matrix and evaluate the performance of the new algorithms with precision and the F1-score. Experimental comparisons are conducted on support vector machine (SVM) and naive Bayes (NB) classifiers. The experimental results demonstrate that the feature selection algorithms proposed in this paper improve performance across various news corpora, although the best feature selection schemes for each type of corpus are different. Further studies of the application of the improved feature selection methods in other languages and the improvement in classifiers are suggested.

Linguistic purification of violence in the press: a comparative critical discourse analysis of execution news

The aim of this study is to probe for the discursive construction of shared beliefs about capital punishment. Drawing upon a specialized corpus of execution news, the study explores representational practices, including social actor naming, action description, and incorporation of voices, and compares respective lexico-grammatical choices that are orchestrated to reconstruct the events of executions in Thailand from 1997 to 2017. The findings reveal that these discursive choices purify state killing and hide the violence of state killing. While representation choices become less overtly negative over time, the underlying ideological position of the press has not changed. Throughout these periods, executions are represented as a routine procedure by minimizing and/or concealing state agency and as a commensurate punishment for the executed individuals, while the law enforcement and justice systems in place are intertextually framed as unproblematic and effective in deterring crime, based on the perspectives of few elite sources. By way of demystifying subtle linguistic choices, this study makes transparent the way in which the press endorses, normalizes, and perpetuates favorable stances toward death penalty in a democratic society.

Tibetan text classification based on graph convolutional network

In this study of Tibetan text classification based on graph convolutional networks, the text graph is first constructed for the Tibetan news corpus based on word-word relations and word-document relations using the constructive relations of Tibetan words and documents, then initialized using the unique thermal representations of words and documents. The embeddings of words and documents are learned using graph convolutional networks under the supervision of the training set of documents, and finally, the text classification problem is transformed into a node classification problem. The graph convolutional network achieves an accuracy of 68.45 % on the Tibetan text classification task, which is higher than other baseline models. In this paper, two comparative experiments are designed to illustrate the effectiveness of graph convolutional networks on Tibetan text classification, which helps people to solve the problem of confusing information in Tibetan texts.

Public communication of technoscience in the news: A cross-linguistic Multidimensional analysis of English and Italian newspapers

Technoscience has historically been endowed by modern societies with cultural authority, which has undergone significant questioning since the end of WWII. In this study, we consider technoscience communication in online newspapers to test whether its linguistic and communicative features differentiate it from other types of news, possibly reflecting greater cultural authority. Focusing on news in English versus Italian, we apply Multidimensional analysis to a comparable English-Italian online news corpus comparing technoscience-related articles with all remaining articles. Results indicate overall linguistic and communicative homogeneity between science and other domains. Nevertheless, small differences seem to suggest an effort to make technoscience accessible and engaging in English, whereas Italian technoscience-related news tends to be more formal and richer in specialised information.

Atrició fonològica i desestabilització de l’estàndard en la televisió valenciana


 
 
 In the context of the Valencian linguistic conflict, the native language undergoes a progressive reduction of the phonological repertoire that tends to eliminate any phoneme that does not exist in the dominant language. This ongoing process exceeds the traits and territory of the well-known apitxat. This adaptation of the phonological system to that of Spanish is accompanied by lexical impoverishment, semantic and grammatical calque of the dominant language. In such a situation, mass media can be a tool not only to increase language use, but also for standardization, that is, the transmission of a normative and genuine model of language. In this contribution, we analyse a corpus of news from the Valencian public television À Punt, in a period of great prominence of the COVID-19 pandemic, to find out what impact the attrition of phonemes has on the speech of professional broadcasters. From a sociophonic point of view, we focus on the pronunciation diverging from the normative Valencian, especially when caused by the approach to Spanish, in order to determine the model of language that this broadcaster transmits to the Valencian society.
 
 

Ontological knowledge inferring approach: Introducing Directed Collocations (DC) and Joined Directed Collocations (JDC)

The growing need of utilizing unstructured knowledge embedded in open-domain natural language text into machine-processable forms requires the induction of hardly extracted structured knowledge into knowledge bases which makes the Semantic Web vision a reality. In this context, ontologies, and ontological knowledge (triples) plays a vital role. This research introduces two novel concepts named Directed Collocation (DC) and Joined Directed Collocation (JDC) along with a methodical application of them to infer new ontological knowledge. Introduced Quality-Threshold-Value (QTV) parameter improves the quality of the inferred ontological knowledge. Having set a moderate value (3) for QTV, this approach inferred 95,491 new ontological knowledge from 43,100 triples of open domain Sri Lankan English news corpus. Indeed, the outcome was approximately doubled in size as the source corpus. Some inferred ontological knowledge was identical with the original corpus content, which evidences the accuracy of this approach. The remaining were validated using inter-rater agreement method (high reliability) and out of which around 56% were estimated as effective. The inferred outcome which is in the triple format may use in any knowledge base. The proposed approach is domain independent. Thus, helps to construct/extend ontologies for any domain with the help of less or no human specialists.