Hasil untuk "Discourse analysis"

Keir Elam

K. Wales

Nur Hafazah Sharin, Mira Kartiwi

Technical and Vocational Education and Training (TVET) has become a key priority for the Malaysian government to enhance the system, better aligning it with industrial demands and workforce needs. The primary priority is to ensure that students and graduates acquire in-demand skills, thereby increasing their employability and creating more attractive job opportunities. Due to rapid technological advancements, social media has emerged as a powerful platform for public discourse where discussions on TVET programs, policies, and perceptions occur extensively. Among these platforms, Facebook is a widely used space for public interactions through posts and comments. This study employs sentiment analysis to analyse TVET-related discussions on Facebook, categorising sentiment into positive, neutral, and negative polarities. The Term Frequency-Inverse Document Frequency (TF-IDF) method is utilised to extract meaningful insights, and six classifiers, comprised of Support Vector Machine (SVM), Naïve Bayes (NB), Decision Tree (DT), Random Forest (RF), K-Nearest Neighbour (KNN), and Logistic Regression (LR), are applied. Using an 80%-20% training and testing split, results indicate that SVM achieves the highest accuracy performance, with a score of 0.62, outperforming other classifiers. Hence, this study provides valuable insights for policymakers and relevant stakeholders in the TVET ecosystem. By leveraging sentiment analysis and machine learning, decision-makers can better understand public perceptions and develop well-informed strategies to realign and enhance the TVET system. ABSTRAK: Pendidikan dan Latihan Teknikal dan Vokasional (TVET) menjadi keutamaan kerajaan Malaysia bagi meningkatkan sistem agar lebih selaras dengan permintaan industri dan keperluan tenaga kerja.  Keutamaan ini adalah bagi memastikan pelajar dan graduan memperoleh kemahiran yang diperlukan, meningkatkan kebolehpekerjaan serta mewujudkan lebih banyak peluang pekerjaan. Kepesatan kemajuan teknologi menyebabkan media sosial muncul sebagai platfom berpengaruh bagi wacana awam di mana perbincangan mengenai program, dasar, dan persepsi TVET berlangsung secara meluas. Antara platfom tersebut, Facebook menjadi medium terbanyak digunakan bagi interaksi awam melalui hantaran dan komen.  Kajian ini menggunakan analisis sentimen bagi menganalisis perbincangan berkaitan TVET di Facebook dengan mengkategorikan sentimen kepada positif, neutral, dan negatif. Kaedah Frekuensi Dokumen Terma Frequency-Inverse (TF-IDF) digunakan bagi mengekstrak pandangan bermakna dan seterusnya menerapkan enam pengklasifikasi yang terdiri daripada Mesin Sokongan Vaktor (SVM), Naïve Bayes (NB), Pokok Keputusan (DT), Rawak Forest (RF), K-Nearest Neighbour (KNN), dan Regriasi Logistik (LR). Menggunakan peratusan data pembahagian latihan dan ujian sebanyak 80%-20%, dapatan kajian menunjukkan bahawa SVM mencapai prestasi ketepatan tertinggi dengan skor 0.62,  mengatasi pengklasifikasi lain. Oleh itu, kajian ini memberi pandangan berharga kepada penggubal dasar dan pihak berkepentingan dalam ekosistem TVET. Dengan memanfaatkan analisis sentimen dan pembelajaran mesin, penggubal dasar dapat memperoleh pemahaman mendalam tentang persepsi awam dan membangunkan strategi berinformasi bagi menyelaras dan meningkatkan sistem TVET.

Cătălin Peptan

The conflict that erupted on October 7, 2023, in the Gaza Strip has triggered one of the most severe humanitarian crises in the recent history of the Middle East, leading to massive forced displacements and a sharp deterioration in living conditions for the civilian population. This study examines the interdependencies between forced migration, the process of securitization, and the violation of fundamental human rights through the theoretical framework of the Copenhagen School. By employing a qualitative analysis of international documents (UN, UNHCR, UNRWA, OCHA, Amnesty International, HRW), the research highlights how political and institutional discourses transform Palestinian migration from a humanitarian issue into a matter of security. In this context, the right to asylum, the principle of non-refoulement, and freedom of movement are often suspended under the pretext of safeguarding national security. The ceasefire of October 10, 2025 - mediated by the United States, Egypt, Qatar, and Turkey and at the time this study was completed (October 15, 2025) - marked a pivotal moment, offering a brief diplomatic respite and a framework for reassessing the international response. However, its effects remain limited, reflecting a temporary desecuritization of the discourse without structural changes in the protection mechanisms. The study proposes directions for the desecuritization of migration and for strengthening the legal and institutional instruments aimed at protecting Palestinian refugees within a geopolitical environment characterized by instability and impunity.

Limor Ziv, Gal Yavetz

In this study, we explored political participation among young adults in Israel, with a particular focus on the impact of polarization on their engagement in online discourse. To this end, by extracting a total of 23,223 Facebook activities (e.g., likes, comments) from 50 participants, we analyzed 2,323 partisan-political and 1,434 socio-political patterns in political and social identity expressions. In contrast to previous research that predominantly relied only on self-report or observed activities, we used direct activity logs, combined with in-depth interviews, thereby obtaining a more detailed insight into user behavior. The results revealed that Facebook usage frequently reflects a hidden “political self,” characterized by non-committal interactions (e.g., frequent likes). The study participants were more at ease within the socio-political domain than in partisan-political discussions. We also found that young adults use social media to cautiously navigate their political and social identities in polarized settings. Based on the results, we propose a novel framework for the analysis of social media engagement that could be useful for policymakers, politicians, and social organizations in crafting strategies to target young adults. The study concludes with the discussion of the necessity to confront polarization by promoting a balanced approach to digital discourse among young adults.

Syukri Rizki, Yusrizal Yusrizal

Arabic, regarded as the language of the educated in both present and past Aceh, facilitates the comprehension of numerous Islamic didactic texts authored by scholars from the Arab world and other Muslim regions. A widely studied text connecting the Acehnese people to this scholarship is Jurjānī’s manual, designed to introduce basic grammatical rules to Arabic beginners. Toward the end of the twentieth century, the prolific Acehnese scholar Abu Teupin Raya (Teungku Muḥammad ʻAlī Irsyād) translated Jurjānī’s text into the Acehnese language under the title Farādīs al-Jinān fī Tarjamah al-’Awāmil al-Jurjānī bi Lisān al-’Ulamā’ al-Qudamā’ bil-Āsyī, aiming to simplify Arabic grammar for Acehnese students. This paper explored into the oral features of the text, examining the constants and patterns in the relationship of Arabic, Malay, and Acehnese languages. It highlighted the authors’ method of using Acehnese language to render the original Arabic source and discussed the exclusivity of Acehnese as an official language within the dayah (traditional Islamic schools). Using the Faircloughian Critical Discourse Analysis (CDA) approach, the text was analyzed as discourse fundamentally linked to its socio-political context. The study employed a three-layer interpretation (micro, meso, and macro levels), as suggested by the approach. The findings revealed that the literal translation technique employed by Abu Teupin Raya reflects an oral translation mode commonly practiced by teachers in Acehnese traditional Islamic schools. Furthermore, the awareness of Aceh as a distinct nation was already evident during the authors’ lifetime, highlighting the linguistic distance between Acehnese and other languages, especially Arabic and Malay.

Elisabetta Zurru

A corpus of nineteen posts collected from the social media platform X was analysed using an integrated theoretical and methodological approach. By combining ecostylistics with ecolinguistics, multimodal critical discourse analysis and multimodal studies, the study investigates the online communicative practices of the grassroots environmental movement #SaveBuxwahaForest, in order to gauge their stylistic traits and communicative strategies, functions and effects. The contribution will also explore the link between environmental activism and protection of indigenous peoples’ rights. The analysis shows that background knowledge about (post)colonial policies towards – or against – indigenous peoples in the Indian subcontinent is necessary to fully decode more than one text in the corpus; that verbal and non-verbal figurative language and non-conventional oppositions are major stylistic traits in the digital communication of this environmental movement; that engagement, mobilisation and persuasion are their main communicative functions; and that most of the strategies used are beneficial rather than ambivalent or destructive, as per Stibbe’s ([2015] 2021) classification.

Michelle Amri, Jan Filart, Jesse B Bump

Global health, as noted in the emerging decolonizing global health literature, is built on power asymmetries and inequities, is centred on individuals and organizations in the global north, and involves a north to south diffusion of ideas and resources. Despite increasing attention paid to the decolonization of global health, there is no universal understanding of what this entails, or what associated agenda(s) may be. We argue that decolonizing global health is not possible without interrogating its many power asymmetries. In this article we demonstrate one example, using a critical discourse analysis of a tremendously influential document, the final report of the World Health Organization's Commission on Social Determinants of Health, Closing the gap in a generation: Health equity through action on the social determinants of health. This report brought mainstream attention to health inequities and the broader forces that underpin them. We reasoned that a flagship report focused on equity and the social determinants of health would be sensitive to the many power inequities in global health. Our critical discourse analysis reveals normative views that presume inequity, such as Euro-American-centricity and portraying countries of the global south as behind or inferior to those of the global north and requiring support. Also, we find that many country comparisons exclude rich countries, which hides the full extent of global inequity. By drawing attention to the inequities presumed in language, we illuminate the persistence of neocolonial ideas that accept rather than contest unfairness.

Maxwell P. Opoku, Negmeldin Alsheikh, Daniel Miezah et al.

Background: Although trauma is one of the leading causes of behaviour problems among children with disabilities, there has been limited scholarly interest in trauma management within the discourse of implementation of inclusive education. Objectives: The Substance Abuse and Mental Health Services Administration (SAMHSA) trauma management model was used to study teachers’ awareness of trauma management among students with disabilities studying in regular classrooms. Method: A total of 271 teachers were recruited from two municipalities in the central region of Ghana to complete the Teacher Trauma Management Scale developed for this study. The data were analysed using confirmatory factor analysis, mean scores, multivariate analysis of variances, and linear regression. Results: The results showed teachers’ uncertainty towards trauma management, and a positive correlation was also found between the tenets of the study framework. Conclusion: The study concluded with a recommendation for contextual development of the curriculum to guide teacher training in trauma management. Contribution: Studies on trauma management within the discourse of implementation of inclusive education are scarce. This study extends the literature on inclusive education to teacher development to support trauma management among students with disabilities in regular schools.

Romina Singh, Kenel Keneshwar Singh, Sakul Kundra

This paper investigates the negative and positive politeness strategies used in the syntactic structures and hedging devices in conversation in Bua Fiji Hindi. Studies on politeness in different cultures have received the attention of anthropologists and linguists. This research was based on Brown and Levinson’s, Grice’s and Lakoff’s models of politeness strategies. This work is innovative since politeness has not been studied in Fiji Hindi, and is based on recorded natural discourse. Thirty discussions were recorded and translated. The information was coded into several sentence-level classes and, after that examined. In this paper, analysis of six types of hedging devices in the conversation will be presented – subjectivity markers, performative hedges, clausal mitigators, downgraders, proverbs and politeness maxims. Together with this, syntactic strategies such as honorificity in pronominal use, plurality in verbs, use of particles, and syntactic structures in requests will also be presented. The most significant difference found in the analyses is that females tend to use subjectivity markers predominantly. The research however, did not reveal any gender differences in using syntactic strategies. This research may inspire some new ideas concerning politeness strategies across different cultures in order to understand how cultural differences play a role in people’s politeness behaviour in conversations.

Makhortova, Varvara Александровна, Kutyeva, Marina Viktorovna

The paper examines a complex of figurative meanings of the entomonym “abelha” (“bee”), manifested in the space of the Portuguese poetic text of the 19th–20th centuries. The purpose of the work is to trace the transformation of the image and its semantic increments. We used methods of contextual, stylistic, semantic analysis and linguistic and cultural commentary. The material includes poems by Almeida Garret, Guerra Junqueiro, Antonio Nobre, Fernando Pessoa, Sofi a de Mello, José Saramago, Fernando Echevarría, etc. The bee is one of the few insects that are evaluated mostly positively in the Portuguese language worldview. However, specific individual authors’ figurative associations diff er in their originality, significantly expanding and complementing the corresponding concept. So, a swarm of bees symbolizes both a sensibly organized collective work, and a gang of naughty schoolboys (G. Junqueiro). At the same time, a separate bee, along with diligence in work, is associated with a thirst for life and creativity (S. de Mello), the birth of rhymes swarming in the mind, boiling thoughts (J. Saramago) and the fervor of love (J. Saramago, A. Garret). The fl ight of a bee over blooming gardens is often likened in Portuguese poetry to the craving of a man for a woman, and the bees themselves become an allegory of sparkling glances (E. de Castro), also of lips and kisses – thanks to the association with honey. Sounds’ similes are important as well: they are guitar chords, fado melodies (A. Nobre). The vast variability of poetic interpretations of the bee’s image is cemented into a single whole by the idea of intensity: diligent collective work, ardent love, a passionate desire to live and create (thirst for life), a powerful creative upsurge, irrepressible frolic of children, deep and rich (thick) sounds of fado in silence, the inviting shine of bright eyes. The image of a bee, reinterpreted in Portuguese poems, is distinguished by a variety of associative foundations, originality and versatility. Having undergone significant transformations in poetic discourse, the updated semantic palette of this entomonym enriches the imagery of the Portuguese language.

R. Safitri

The fact of the analysis of discourse according to Poerwadarminta (in Bayardi, 2002:1), discourse is reviewed from the suggestion Vacana ' recitation ' in Sanskrit. Then Vacana entered into the old Javanese language and the new Javanese language became the discourse and the discourse ' speech and speech ', then absorbed into the Indonesian language into a discourse of ' speech or conversation '. So that the discourse is a word or utterance, say or speech. Speech in both oral and written form.

Gábor Bíró, Gergely Gábor Barnaföldi, Péter Lévai

High-energy physics (HEP) provides ever-growing amount of data. To analyse these, continuously-evolving computational power is required in parallel by extending the storage capacity. Such developments play key roles in the future of this field however, these can be achieved also by optimization of existing IT resources. One of the main computing capacity consumers in the HEP software workflow are detector simulation and data analysis. To optimize the resource requirements for these aims, the concept of a dedicated Analysis Facility (AF) for Run 3 has been suggested by the ALICE experiment at CERN. These AFs are special computing centres with a combination of CPU and fast interconnected disk storage modules, allowing for rapid turnaround of analysis tasks on a dedicated subset of data. This in turn allows for optimization of the analysis process and the codes before the analysis is performed on the large data samples on the Worldwide LHC Computing Grid. In this paper, the structure and the progress summary of the Wigner Analysis Facility (Wigner AF) is presented for the period 2020-2022.

D. Schum

Jiaxin Pei, Vítor Silva, Maarten Bos et al.

We propose MINT, a new Multilingual INTimacy analysis dataset covering 13,372 tweets in 10 languages including English, French, Spanish, Italian, Portuguese, Korean, Dutch, Chinese, Hindi, and Arabic. We benchmarked a list of popular multilingual pre-trained language models. The dataset is released along with the SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis (https://sites.google.com/umich.edu/semeval-2023-tweet-intimacy).

Shichao Kan, Zhiquan He, Yigang Cen et al.

Recent methods for deep metric learning have been focusing on designing different contrastive loss functions between positive and negative pairs of samples so that the learned feature embedding is able to pull positive samples of the same class closer and push negative samples from different classes away from each other. In this work, we recognize that there is a significant semantic gap between features at the intermediate feature layer and class labels at the final output layer. To bridge this gap, we develop a contrastive Bayesian analysis to characterize and model the posterior probabilities of image labels conditioned by their features similarity in a contrastive learning setting. This contrastive Bayesian analysis leads to a new loss function for deep metric learning. To improve the generalization capability of the proposed method onto new classes, we further extend the contrastive Bayesian loss with a metric variance constraint. Our experimental results and ablation studies demonstrate that the proposed contrastive Bayesian metric learning method significantly improves the performance of deep metric learning in both supervised and pseudo-supervised scenarios, outperforming existing methods by a large margin.

Rolf Sören Kraußhar, Anastasiia Legatiuk, Dmitrii Legatiuk

In recent years, there is a growing interest in the studying octonions, which are 8-dimensional hypercomplex numbers forming the biggest normed division algebras over the real numbers. In particular, various tools of the classical complex function theory have been extended to the octonionic setting in recent years. However not so many results related to a discrete octonionic analysis, which is relevant for various applications in quantum mechanics, have been presented so far. Therefore, in this paper, we present first ideas towards discrete octonionic analysis. In particular, we discuss several approaches to a discretisation of octonionic analysis and present several discrete octonionic Stokes' formulae.

Mohammad Khalooei, Mohammad Mehdi Homayounpour, Maryam Amirmazlaghani

Deep neural network models are used today in various applications of artificial intelligence, the strengthening of which, in the face of adversarial attacks is of particular importance. An appropriate solution to adversarial attacks is adversarial training, which reaches a trade-off between robustness and generalization. This paper introduces a novel framework (Layer Sustainability Analysis (LSA)) for the analysis of layer vulnerability in an arbitrary neural network in the scenario of adversarial attacks. LSA can be a helpful toolkit to assess deep neural networks and to extend the adversarial training approaches towards improving the sustainability of model layers via layer monitoring and analysis. The LSA framework identifies a list of Most Vulnerable Layers (MVL list) of the given network. The relative error, as a comparison measure, is used to evaluate representation sustainability of each layer against adversarial inputs. The proposed approach for obtaining robust neural networks to fend off adversarial attacks is based on a layer-wise regularization (LR) over LSA proposal(s) for adversarial training (AT); i.e. the AT-LR procedure. AT-LR could be used with any benchmark adversarial attack to reduce the vulnerability of network layers and to improve conventional adversarial training approaches. The proposed idea performs well theoretically and experimentally for state-of-the-art multilayer perceptron and convolutional neural network architectures. Compared with the AT-LR and its corresponding base adversarial training, the classification accuracy of more significant perturbations increased by 16.35%, 21.79%, and 10.730% on Moon, MNIST, and CIFAR-10 benchmark datasets, respectively. The LSA framework is available and published at https://github.com/khalooei/LSA.

Amanda Siebert-Evenstone, Golnaz Arastoopour, W. Collier et al.

Analyses of learning based on student discourse need to account not only for the content of the utterances but also for the ways in which students make connections across turns of talk. This requires segmentation of discourse data to define when connections are likely to be meaningful. In this paper, we present an approach to segmenting data for the purposes of modeling connections in discourse using epistemic network analysis. Specifically, we use epistemic network analysis to model connections in student discourse using a temporal segmentation method adapted from recent work in the learning sciences. We compare the results of this study to a purely conversation-based segmentation method to examine the affordances of temporal segmentation for modeling connections in discourse.

Jeannett Martin, R. Veel