News Classification Research Articles

How does the development of COVID-19 affect the public's engagement to fake news rebuttal microblogs?

COVID-19 infodemics are intertwined with the dynamic development of the pandemic. During infodemics, the challenges and difficulties are to reveal the dynamic evolution of public’s engagement to fake news rebuttals information and break through the bottleneck of previous static research. Motivated by this, fake news rebuttals engagement index (FNREI) is proposed to state the public engagement towards COVID-19 related fake news rebuttals. A total of 19,603 fake news rebuttals microblogs from Weibo are divided into four categories by latent dirichlet allocation (LDA): “status and confirmed cases of COVID-19″, ”transmission mode and living conditions of coronavirus“, ”prevention and detection of COVID-19″ and “overseas fake news and derivative events”. Time-varying parameter vector auto-regressive model (TVP-VAR) is used to explore the time-varying impact of the COVID-19 life cycle and the appearance of virus mutation (Delta, Omicron) on four categories of FNREI time series. The research results suggest that under the impact of different major events during the COVID-19 life cycle, the public's engagement to different categories of COVID-19 related fake news has different priorities, its duration and intensity is dynamically changing. This study explores the evolution of individuals’ engagement to fake news rebuttals information during the COVID-19 and provides a basis for improving fake news governance effectiveness.

News Alert App

The News Alert Android App is a revolutionary mobile application designed to keep users informed with real- time access to breaking news from across the globe. With its user-friendly interface, the app ensures that users can easily navigate through various news categories and receive instant notifications for critical updates and emergencies. Offering personalized news preferences, users can customize their experience by selecting specific topics, regions, or industries of interest. Beyond text-based news, the app provides multimedia content, including images and videos, to enhance the storytelling experience. Social media integration allows users to share news updates with their networks, encouraging meaningful discussions and raising awareness about important issues. With AI-driven algorithms, the app delivers smart notifications to avoid overwhelming users with redundant or irrelevant information. Whether online or offline, users can access up-to-date news, bookmark articles, and stay connected to the world's events. The News Alert Android App is a powerful tool for those seeking timely and reliable information, fostering an informed and engaged community in today's fast-paced information age. The news alert app includes modules like home, science technology, sports, health, entertainment for users. Keyword: Real-Time News Delivery, User Experience, User- Centric Design, Push Notification.

The impact of news on South African sovereign bond yields

ABSTRACT A reverse event study approach was used to investigate how the South African sovereign bond yield curve reacts to headline news. The change in daily yields, calculated as the difference between the natural log of zero-coupon yields on consecutive business days, were used in the analysis. Dates of abnormal daily yield changes were identified using GARCH models. News items for the sample period were classified into categories using supervised machine learning. A regression model was fitted to determine the link between the abnormal yield changes and the news categories. The results indicated that, for abnormal increases in yield (negative news), political news had an impact on all nodes. For abnormal decreases in yield (positive news), economic news had the greatest impact on the 10-year and political news on the 15- and 20-year nodes of the yield curve.

Unifying Sentence Transformer Embedding and Softmax Voting Ensemble for Accurate News Category Prediction

The study focuses on news category prediction and investigates the performance of sentence embedding of four transformer models (BERT, RoBERTa, MPNet, and T5) and their variants as feature vectors when combined with Softmax and Random Forest using two accessible news datasets from Kaggle. The data are stratified into train and test sets to ensure equal representation of each category. Word embeddings are generated using transformer models, with the last hidden layer selected as the embedding. Mean pooling calculates a single vector representation called sentence embedding, capturing the overall meaning of the news article. The performance of Softmax and Random Forest, as well as the soft voting of both, is evaluated using evaluation measures such as accuracy, F1 score, precision, and recall. The study also contributes by evaluating the performance of Softmax and Random Forest individually. The macro-average F1 score is calculated to compare the performance of different transformer embeddings in the same experimental settings. The experiments reveal that MPNet versions v1 and v3 achieve the highest F1 score of 97.7% when combined with Random Forest, while T5 Large embedding achieves the highest F1 score of 98.2% when used with Softmax regression. MPNet v1 performs exceptionally well when used in the voting classifier, obtaining an impressive F1 score of 98.6%. In conclusion, the experiments validate the superiority of certain transformer models, such as MPNet v1, MPNet v3, and DistilRoBERTa, when used to calculate sentence embeddings within the Random Forest framework. The results also highlight the promising performance of T5 Large and RoBERTa Large in voting of Softmax regression and Random Forest. The voting classifier, employing transformer embeddings and ensemble learning techniques, consistently outperforms other baselines and individual algorithms. These findings emphasize the effectiveness of the voting classifier with transformer embeddings in achieving accurate and reliable predictions for news category classification tasks.

Dezinformacja. Próba ujęcia dyskursywnego

The categories of disinformation and fake news are among the more frequently invoked concepts in the context of discussions about modernity. On the one hand, there is widespread agreement on the special (negative) role of media phenomena defined in this way, while on the other hand, a lot of research is being conducted to shed light on their nature and principles of operation. Meanwhile, there is a fundamental – at least from the point of view of the social sciences – problem with these concepts. Namely, we do not have their precise definitions. The presented text has the task of analyzing the state of affairs in this regard and proposing an adequate approach. This approach, in fact, identifies the category of disinformation with the category of fake news (in the broader sense), while pointing out the constitutive features of both these phenomena, related to the categories of truth, intention and media format. These features are analyzed in the context of the category of the media genre of information (more specifically, the information function understood in a pragmatic way) and the discursive framework of journalism, into which, as will be shown in the text, disinformation attempts to intrude.

Automated Arabic News Classification using the Convolutional Neural Network

Automated Arabic News Classification using the Convolutional Neural Network

Sistem Klasifikasi Berita Menggunakan Metode Text Mining pada Website Pusat Kegiatan Belajar Masyarakat

The use of information systems via websites is not uncommon in this current era. This is because the importance of information quality can affect trust, credibility of an organization, and often used as a promotional media. However, the problem arises when there are increasing numbers of news to be informed, which becomes a problem for web managers. Therefore, a faster method and proper news classification system is needed to avoid future problems. Thus, this research uses the text mining method and pure Term Frequency algorithm to calculate the weight of each word, in order to determine which category the news belongs to automatically. To simplify the system design process, Unified Modeling Language (UML) and PIECES analysis are used to analyze the impact factors that will arise later. Based on the results of the classification system testing, it has been able to provide solutions to categorize information in PKBM, even though there are many news articles with different categories.

Food safety news events classification via a hierarchical transformer model

In light of the significance of regulatory authorities and the rising demand for information disclosure, a vast amount of information on food safety news reports is readily accessible on the Internet. The extraction of such information for precise classification and provision of appropriate safety alerts based on their respective categories has emerged as a challenging problem for academic research. Given that most food safety-related events in news reports comprise lengthy text, the pre-trained language models currently employed for text analysis are generally limited in their capability to handle long documents. This paper proposes a long-text classification model utilising hierarchical Transformers. We categorise information in long documents into two distinct types: (1) multiple text chunks meeting the length constraint and (2) essential sentences within long documents, such as headings, paragraph start and end sentences, etc. Initially, our proposed model utilises the text chunks as input to the BERT model. Then, it concatenates the output of the BERT model with the important sentences from the document and use them as input to the Transformer model for feature transformation. Finally, we utilise a classifier for food safety news classification. We conducted several comparative experiments with various commonly used text classification models on a dataset constructed from publicly available information on food regulatory websites. Our proposed method outperforms existing methods, establishing itself as the leading approach in terms of performance.

Optimal Number Data Trains in Hoax News Detection of Indonesian using SVM and Word2Vec

Along with the development of the era of technological development also has an increase. Information dissemination occurs very quickly on social media, especially Twitter. On Twitter, only some news circulating is necessarily accurate information. Lots of information that is spread is hoax news that irresponsible individuals apply. In this research, the author will build a system to determine the optimal amount of data trained in the hoax news classification process. In this study, the authors will use the support vector machine and word2vec algorithms to classify hoax and non-hoax news on the system to be created. In this study, five experiments were carried out with the number of train data used as many as 5000, 10000, 15000, 20000, and 25000. 5000 data train results in an accuracy of 77.28%, 10000 data train produce an accuracy of 79.68%, data 15,000 trains produce an accuracy of 79.892%, 20,000 data trains produce an accuracy of 80,416%, and 25,000 data trains produce an accuracy of 81,184%, by using a combination of unigram with token full token selection. This research aims to build a hoax detection system that can determine the optimal amount of data training to use. Also, this research is used to see the performance of the Support Vector Machine algorithm with Word2Vec in detecting hoax news

Rekomendasi Berita Berkaitan dengan Menerapkan Algoritma Text Mining dan TF-IDF

News presentation is generally structured in such a way that the information presented is well grouped, but the use of electronic media does not necessarily offer complete news categories because not all of the space offered can be filled with good presentation, so special treatment is needed so that readers get the news. needed which is arranged based on recommendations. To arrange this research to be more structured, the authors carried out several stages in completing the research, namely the Problem Identification Stage, Literature Study Stage, Data Collection Stage, Text Mining and TF-IDF Algorithm Implementation Stage, and conclusions. The author implements the text mining and TF-IDF algorithms in processing news title data starting with the Text Mining Algorithm where this stage is a preprocessing stage with the aim that the data to be processed is a basic word so that the weighting process in the TF-IDF Algorithm is not too broad. After the text mining stage, it will proceed to the TF-IDF stage, namely weighting the terms in each document. Text mining and TF-IDF algorithms are able to provide appropriate news recommendations based on the highest similarity in meaning both in terms of topic and object of the news title, for future research it is recommended to use other algorithms such as cosine similar so that recommendations are not only generated from the suitability of words but can also see the similarity of meaning so that research results can be even better.

A Graph Neural Network Node Classification Application Model with Enhanced Node Association

This study combines the present stage of the node classification problem with the fact that there is frequent noise in the graph structure of the graph convolution calculation, which can lead to the omission of some of the actual edge relations between nodes and the appearance of numerous isolated nodes. In this paper, we propose the graph neural network model ENode-GAT for improving the accuracy of small sample node classification using the method of external referencing of similar word nodes, combined with Graph Convolutional Neural Network (GCN), Graph Attention Network (GAT), and the early stop algorithm. In order to demonstrate the applicability of the model, this paper employs two distinct types of node datasets for its investigations. The first is the Cora dataset, which is widely used in node classification at this time, and the second is a small-sample Stock dataset created by Eastern Fortune’s stock prospectus of the Science and Technology Board (STB). The experimental results demonstrate that the ENode-GAT model proposed in this paper obtains 85.1% classification accuracy on the Cora dataset and 85.3% classification accuracy on the Stock dataset, with certain classification advantages. It also verifies the future applicability of the model to the fields of stock classification, tender document classification, news classification, and government announcement classification, among others.

Arabic News Classification Based on the Country of Origin Using Machine Learning and Deep Learning Techniques

With the rise of Arabic news articles published daily, people are becoming increasingly concerned about following the news from reliable sources, especially regarding events that impact their country. To assess a news article’s significance to the user, it is essential to identify the article’s country of origin. This paper proposes several classification models that categorize Arabic news articles based on their country of origin. The models were developed using comprehensive machine learning and deep learning techniques with several feature training methods. The results show the ability of our model to classify news articles based on their country of origin, with close accuracy between machine learning and deep learning techniques of up to 94%.

Identifying multimodal misinformation leveraging novelty detection and emotion recognition.

With the growing presence of multimodal content on the web, a specific category of fake news is rampant on popular social media outlets. In this category of fake online information, real multimedia contents (images, videos) are used in different but related contexts with manipulated texts to mislead the readers. The presence of seemingly non-manipulated multimedia content reinforces the belief in the associated fabricated textual content. Detecting this category of misleading multimedia fake news is almost impossible without relevance to any prior knowledge. In addition to this, the presence of highly novel and emotion-invoking contents can fuel the rapid dissemination of such fake news. To counter this problem, in this paper, we first introduce a novel multimodal fake news dataset that includes background knowledge (from authenticate sources) of the misleading articles. Second, we design a multimodal framework using Supervised Contrastive Learning (SCL) based novelty detection and Emotion Prediction tasks for fake news detection. We perform extensive experiments to reveal that our proposed model outperforms the state-of-the-art (SOTA) models.

Panic buying and fake news in urban vs. rural England: A case study of twitter during COVID-19

This paper explores the potential association between the spread of fake news and the panic buying behavior, in urban and rural UK, widely accessible on Twitter since COVID 19 was announced by the WHO as a global pandemic. It describes how consumer's behavior is affected by the content generated over social media and discuss various means to control such occurrence that results in an undesirable social change. The research methodology is based on extracting data from texts on the subject of panic buying and analysing both the total volume and the rate of fake news classification during COVID-19, through crowdsourcing techniques with text-mining and Natural Language Processing models. In this paper, we have extracted the main topics in different phases of the pandemic using term frequency strategies and word clouds as well as applied artificial intelligence in exploring the reliability behind online written text on Twitter. The findings of the research indicate an association between the pattern of panic buying behavior and the spread of fake news among urban and rural UK. We have highlighted the magnitude of the undesired behavior of panic buying and the spread of fake news in the rural UK in comparison with the urban UK.

Detecting fake news with deep neural networks

The raise of social media and the easy access to obtain and spread news has increased the probability of various news dissemination. However, an undetermined factor in usage scenario is the authenticity of spreading news, which has posed detrimental effects on individuals and society in current scenario. It is imperative to develop proper mechanism for fake news detection. These years, with the sharp development of deep learning, referring to the method of training multi-layer artificial neural networks, some researchers have utilized deep learning in fake news detection and classification. Although many studies have achieved a preeminent level and high accuracy, scarcity of review work regarding fake news detection with deep learning is still an important problem. This work urges to derive a comprehensive review of fake news detection and classification techniques for further study in the aspect, and mechanism in various studies for sake of raising accuracy. This work also figure out some possible challenges with existing limitation based on previous research.

Framing News About Turkey in Albanian and Kosovo Online Newspapers: The Case of Lake Memorial About 2016 Coup Attempt in Turkey

This research examines the language used in media articles with framing content against Turkey in the Balkans, specifically in Albania and Kosovo. Albanian national identity and the nationalism that arose following the breakup of the Ottoman Empire have periodically resurfaced throughout modern history. Today, Albanian nationalism can sever ties with modern Turkey, often portraying the past and present in a negative light. The study focuses on how Albanian media with a nationalist perspective covered the construction of a Turkish memorial in Tirana's park, which was intended as a symbol of friendship between the two countries and as a tribute to the martyrs of July 15th. A total of 69 news articles were analyzed using both quantitative and qualitative methods, and categories of news framing related to a nationalist perspective in the journalists' reporting were identified. The results showed that media outlets in both countries covered the event, but Kosovo media displayed a higher level of framing against Turkey, despite the fact that the event took place in Albania. This nationalist approach suggests that Turkey continues to be perceived as a geostrategic, political, and cultural threat to the Albanian nation in the Balkans.

TwIdw—A Novel Method for Feature Extraction from Unstructured Texts

This research proposes a novel technique for fake news classification using natural language processing (NLP) methods. The proposed technique, TwIdw (Term weight–inverse document weight), is used for feature extraction and is based on TfIdf, with the term frequencies replaced by the depth of the words in documents. The effectiveness of the TwIdw technique is compared to another feature extraction method—basic TfIdf. Classification models were created using the random forest and feedforward neural networks, and within those, three different datasets were used. The feedforward neural network method with the KaiDMML dataset showed an increase in accuracy of up to 3.9%. The random forest method with TwIdw was not as successful as the neural network method and only showed an increase in accuracy with the KaiDMML dataset (1%). The feedforward neural network, on the other hand, showed an increase in accuracy with the TwIdw technique for all datasets. Precision and recall measures also confirmed good results, particularly for the neural network method. The TwIdw technique has the potential to be used in various NLP applications, including fake news classification and other NLP classification problems.

Indonesian news classification application with named entity recognition approach

Nowadays, many netizens search for news via search engines with countless amounts of information, so it is increasingly difficult to determine when the number of news articles that appear changes very quickly and dynamically. Thus, it is necessary to process the extraction of news information to display the core information of the news. Problems arise, especially in Indonesian, which has a structure of various noun phrase entities with shallow parsing or grammatical induction. Named Entity Recognition (NER) has the opportunity to overcome this because it can extract news entities in depth, starting from proper nouns in text documents containing information search, machine translation, answering questions, and automatic summarization. This study aims to apply NER in Indonesian language news classification. This study uses Design-Based Research whose process includes (1) pre-implementation, (2) design, (3) implementation and revision, and finally, (4) reflection and evaluation. This application was developed on the platform python, streamlit, BeautifulSoup, gnews, and spacy library. The results of application accuracy testing have an F1-score value of 89.69% for all entities consisting of place, figure, day, date, and organization.

CLASSIFICATION OF FAKE NEWS IN UKRAINE AND ABROAD

<p><strong><em>The aim </em></strong><em>of the work is to propose a broad classification of fake news based on the generalization of Ukrainian and international research.</em></p><p><strong><em>Research methodology.</em></strong><em> Both theoretical and empirical research methods were used in the research process. The research methodology consisted of several stages. The first is data collection. This method was used to build a dataset of fake news articles from various sources. These sources included known purveyors of fake news, such as clickbait sites or biased blogs, as well as reputable news sources that have published fake news. The next stage was extraction of fake news features. After collecting a dataset of desinformation materials, we extract relevant functions that can be used as keywords for searching in Google. These data include word frequencies, grammatical structures, or other linguistic features that are known to be associated with fake news.</em></p><p><strong><em>Results.</em></strong><em> Western researchers distinguish ten types of «fake news» [7]. Each of the ten forms of deceptive or illusory content carries a different level of threat, impact, and intent. The focus should be on identifying the types of content that are malicious and pose a threat of panic and confusion. Foreign researchers distinguish the following types of fakes: fake news, manipulation, deep fakes, puppet news, phishing, spreading rumors, bots, disinformation, clickbait, satire and parody. The above classification is quite narrow, as it covers specific examples of fake media publications. Considering that the media market and the Internet as a platform are dynamic, changing and reacting to external factors, a broader classification was proposed that would work in the longer term and that would also be able to adapt to dynamic changes in the genre.</em></p><p><strong><em>Novelty.</em></strong><em> The novelty of this work is the proposed broad classification of fake news in media outlets on the basis of theoretical and empirical research. </em></p><p><strong><em>Practical meaning. </em></strong><em>The obtained information can be used in further monitoring and research of fake news in Ukrainian and international media outlets. By accurately classifying fake news, the audience and journalists can identify the sources of misinformation and track the spread of false information. By developing different tools to classify fake news, other researchers can help educate the public on how to spot false information online and avoid being misled, which is an important aspect of media literacy.</em></p><p><strong><em>Key words:</em></strong><em> fake news, disinformation, media, audience, clickbait.</em></p>

Sentiment Gradient - Improving Sentiment Analysis with Entropy Increase

Information sharing on the Web has also led to the rise and spread of fake news. Considering that fake information is generally written to trigger stronger feelings from the readers than simple facts, sentiment analysis has been widely used to detect fake news. Nevertheless, sarcasm, irony, and even jokes use similarwritten styles, making the distinction between fake and fact harder to catch automatically. We propose a new fake news Classifier that considers a set of language attributes and the gradient of sentiments contained in a message. Sentiment analysis approaches are based on labelling news with a unique value that shrinks the entire message to a single feeling. We take a broader view of a message’s sentiment representation, trying to unravel the gradient of sentiments a message may bring. We tested our approach using two datasets containing texts written in Portuguese: a public one and another we created with more up-to-date news scrapped from the Internet. Although we believe our approach is general, we tested for the Portuguese language. Our results show that the sentiment gradient positively impacts the fake news classification performance with statistical significance. The F-Measure reached 94 %, with our approach surpassing available ones (with a p-value less than 0.05 for our results).