Text Data Research Articles

BERTAgent: The development of a novel tool to quantify agency in textual data.

Pertaining to goal orientation and achievement, agency is a fundamental aspect of human cognition and behavior. Accordingly, detecting and quantifying linguistic encoding of agency are critical for the analysis of human actions, interactions, and social dynamics. Available agency-quantifying computational tools rely on word-counting methods, which typically are insensitive to the semantic context in which the words are used and consequently prone to miscoding, for example, in case of polysemy. Additionally, some currently available tools do not take into account differences in the intensity and directionality of agency. In order to overcome these shortcomings, we present BERTAgent, a novel tool to quantify semantic agency in text. BERTAgent is a computational language model that utilizes the transformers architecture, a popular deep learning approach to natural language processing. BERTAgent was fine-tuned using textual data that were evaluated by human coders with respect to the level of conveyed agency. In four validation studies, BERTAgent exhibits improved convergent and discriminant validity compared to previous solutions. Additionally, the detailed description of BERTAgent's development procedure serves as a tutorial for the advancement of similar tools, providing a blueprint for leveraging the existing lexicographical data sets in conjunction with the deep learning techniques in order to detect and quantify other psychological constructs in textual data. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

MoHGCN: Momentum Hypergraph Convolution Network for Cross-modal Retrieval

Cross-modal retrieval tasks, encompassing the retrieval of image-text, video-audio, and more, are progressively gaining significance in response to the exponential growth of information on the Internet. However, there has always been a cloud hanging over multimodal tasks due to the inherent challenges in aligning different modalities with distinct physical meanings. Most previous works simply rely on a single multimodal encoder or a novel similarity calculation for fusion, which often result in unsatisfactory performance. To tackle this challenge, we introduce a Momentum Hypergraph Convolutional Network (MoHGCN) for multimodal representation learning, which strengthens the alignment of both visual and textual data before the retrieval process. Specifically, MoHGCN utilizes contrastive learning to select the most challenging negative and positive samples to form hyperedges, and completes the modality alignment through two rounds of fusion. Subsequently, the fully integrated node features and global features are fused using a fusion encoder to obtain the final multimodal representation vector for image-text retrieval. Extensive experiments are conducted on two widely-used datasets, namely Flickr30K and MSCOCO, to demonstrate the superiority of the proposed MoHGCN approach in achieving the state-of-the-art performances.

The Analysis of Cultural Content The Eight Grade Junior High School Students English Textbook

The aims of this research are to analyze the culture categories and get the most dominant category presented in the textbook. Qualitative research design, specifically content analysis of textbooks is employed in the study. This research use data analysis according to Miles and Huberman (1992) explain that analysis involves three concurrent activities: data reduction, data display, and conclusion drawing/verification. The primary data source for this study is the researcher employs established categories to analyze the data, including food and drink, locations, art, ceremonies, social behavior, beliefs and values, history, and individuals. These categories were developed by Adoskou, Moran, Dweik, and Al-Sayed in 1990, 2001, and 2015. The textual and visual data are examined based on the cultural categories of source culture, target culture, and worldwide culture. These categories were developed by Cortazzi & Jin in 2022.

The Concept of Community Empowerment versus Community Advancement: A Comparative Study in the Context of Social Development in Indonesia

Community interventions through community empowerment and community advancement are crucial within Indonesia’s social development agenda, yet their effectiveness is often constrained by conceptual ambiguity between these two approaches. This research aimed to conduct an in-depth comparative study of community empowerment versus community advancement within the Indonesian context, encompassing a comparative analysis of concepts and theoretical foundations, a review of program implementation examples, and identifying their application challenges and opportunities. Employing a qualitative approach with a literature study method and secondary document analysis, this study thematically and comparatively analyzed various textual data sources. The analysis revealed fundamental differences between the two concepts: community empowerment inherently focuses on enhancing agency and shifting power relations, while community advancement is more oriented towards improving socio-economic conditions and infrastructure, although conceptual intersections and significant terminological conflation were also identified. In terms of implementation, programs reflecting community empowerment tend to emphasize substantive participation and strengthening internal capacities, whereas advancement-oriented programs focus more on achieving physical outputs; however, hybrid practices and the strong influence of contextual variations are prominent in the field. Both approaches face systemic challenges (e.g., institutional capacity, sustainability) and specific challenges (e.g., elite capture in empowerment, maintenance issues in advancement), yet opportunities for synergy also exist through decentralization frameworks and technology adoption. It is concluded that a clear understanding of the essential differences, similarities, and potential synergies between community empowerment and community advancement forms a fundamental basis for designing and implementing more effective, contextual, adaptive, and transformative social development interventions in Indonesia.

Signal or pressure? Retail investor attention and MD&A quality

PurposeThis study examines the impact of retail investor attention on the quality of management discussion and analysis (MD&A) sections in the annual reports of A-share main board companies listed in China, framed within the context of limited attention theory.Design/methodology/approachThis study analyzes textual data from the MD&A sections of annual reports and interaction data from the investor-firm interaction platforms of the Shanghai and Shenzhen Stock Exchanges in China, covering 2015–2019. It applies the Continuous Bag of Words (CBOW) model and t-SNE dimensionality reduction technique to construct an indicator for MD&A information content. Additionally, the study assesses the moderating effects of factors, including state ownership, media coverage, regulatory inquiry letters and economic policy uncertainty, on the relationship between retail investor attention and MD&A quality.FindingsThe results show that retail investor attention significantly enhances MD&A quality, primarily through the mechanism of signal transmission. This relationship is moderated by factors including state ownership, media coverage, regulatory inquiry letters and economic policy uncertainty. Moreover, the study indicates that the positive impact of retail investor attention on MD&A quality significantly reduces the risk of stock price crashes and improves firm performance.Originality/valueThis study contributes to the literature by providing new insights into the role of retail investor attention in enhancing MD&A disclosure quality, highlighting retail investors as an informal monitoring mechanism that supports capital market stability. It addresses a gap by incorporating limited attention theory into the analysis of retail investor attention and MD&A quality and introduces an innovative approach to measure MD&A information content. The findings offer valuable implications for improving disclosure practices among listed companies.

A Medical Multimodal Large Language Model for Pediatric Pneumonia.

Pediatric pneumonia is the leading cause of death among children under five years worldwide, imposing a substantial burden on affected families. Currently, there are three significant hurdles in diagnosing and treating pediatric pneumonia. Firstly, pediatric pneumonia shares similar symptoms with other respiratory diseases, making rapid and accurate differential diagnosis challenging. Secondly, primary hospitals often lack sufficient medical resources and experienced doctors. Lastly, providing personalized diagnostic reports and treatment recommendations is labor-intensive and time-consuming. To tackle these challenges, we proposed a Medical Multimodal Large Language Model for Pediatric Pneumonia(P2Med-MLLM). It was capable of handling diverse clinical tasks-such as generating free-text medical records and radiology reports-within a unified framework. Specifically, P2Med-MLLM was trained on a large-scale dataset, including real clinical information from 163,999 outpatient and 8,684 inpatient cases. It can process both plain text data(e.g., outpatient and inpatient records) and interleaved image-text pairs(e.g., 2D chest X-ray images, 3D chest Computed Tomography images, and corresponding radiology reports). We designed a three-stage training strategy to enable P2Med-MLLM to comprehend medical knowledge and follow instructions for various clinical decision-support tasks. To rigorously evaluate P2Med-MLLM's performance, we conducted automatic scoring by the large language model and manual scoring by the specialist on the test set of 642 samples, meticulously verified by pediatric pulmonology specialists. The results demonstrated the reliability of automated scoring and the superiority of P2Med-MLLM. This work plays a crucial role in assisting doctors with prompt diagnosis and treatment planning, reducing severe symptom mortality rates, and optimizing the allocation of medical resources.

Web Mining to Detect Online Spread of Terrorism

Abstract: In the recent times, terrorism has grown in an exponential manner in certain parts of the world. This enormous growth in terrorist activities has made it important to stop terrorism and prevent its spread before it causes damage to human life or property. With development in technology, internet has become a medium of spreading terrorism through speeches and videos. Terrorist organizations use the medium of the internet to harm and defame individuals and also promote terrorist activities through web pages that force people to join terrorist organizations and commit crimes on the behalf of those organizations. Web mining and data mining are used simultaneously for the purpose of efficient system development. Web mining even consists of many different text mining methods that can be helpful to scan and extract relevant data from unstructured data. Text mining is very helpful in detecting various patterns, keywords, and significant information in unstructured texts. Data mining and web mining systems are used for mining from text widely. Data mining algorithms are used to manage organized data sets and web mining algorithms can be helpful in mining and extracting from unstructured web pages and text data that is available across the web. Websites built in different platforms have varying data structures and that makes it quite difficult to read for a single algorithm. Keywords: Terrorism, naïve-bayes, random forest, online spread

Automatic construction of risk transmission network about subway construction based on deep learning models

Safety risks management is a critical part during the subway construction. However, conventional methods for risk identification heavily rely on experience from experts and fail to effectively identify the relationship between risk factors and events embedded in accident texts, which fail to provide substantial guidance for subway safety risks management. With a dataset comprising 562 occurrences of subway construction accidents, this study devised a domain-specific entity recognition model for identifying safety hazards during the subway construction. The model was constructed by a Bidirectional Long Short-Term Memory Network with Conditional Random Fields (BiLSTM-CRF). Additionally, a domain-specific entity causal relation extraction model employing Convolutional Neural Networks (CNN) was also developed in thsi model. The constructed models automatically extract safety risk factors, safety events, and their causal relationships from the texts about subway accidents. The precision, recall, and F1 scores of Metro Construction Safety Risk Named Entity Recognition Model (MCSR-NER-Model) all exceeded 77%. Its performance in the specialized domain named entity recognition (NER) with a limited volume of textual data is satisfactory. The Metro Construction Safety Risk Domain Entity Causal Relationship Extraction Model (MCSR-CE-Model) achieved an impressive accuracy, recall, and F1 score of 98.96%, exhibiting excellent performance. Moreover, the extracted entities were normalized and domain dictionary was developed. Based on the processed entities and relationships processed by the domain dictionary, 533 domain entity causal relation triplets were obtained, facilitating the establishment of the directed and unweighted complex network and case database about the risks of subway construction. This research successfully converted accident texts into a causal chain structure of “safety risk factors to risk events,” providing detailed categorization of safety risks and events. Concurrently, it revealed the interrelationships and historical statistical patterns among various safety risk factors and categories of risk events through the complex safety risks network. The construction of the database facilitated project managers in conducting management decisions about safety risks.

DeepSarcasm: A BiLSTM and GloVe Powered Model for Identifying Sarcasm in Context

This project explores the intricate challenge of sarcasm detection in textual data using advanced Natural Language Processing (NLP) techniques. The primary goal is to create a model capable of accurately identifying and classifying sarcastic remarks within various contexts. We address this by leveraging Bidirectional Long Short-Term Memory (BiLSTM) networks, known for their ability to understand context by processing data in both forward and backward directions. To enhance semantic understanding, GloVe embeddings are employed to capture word relationships and contextual nuances. Our methodology encompasses comprehensive data preprocessing steps - such as tokenization, stopword removal, and lemmatization - to ensure clean and coherent text input. The BiLSTM model is trained on a diverse dataset that includes both sarcastic and non-sarcastic text samples, facilitating the learning of distinctive patterns. We evaluate the model’s performance using metrics like accuracy, precision, recall, and F1-score, anticipating that our model will effectively discern subtle sarcastic cues and outperform baseline methods. This study’s results have significant implications for sentiment analysis, social media monitoring, and conversational AI systems. Future directions include extending the model to handle multilingual sarcasm detection, integrating real-time data processing, and addressing ethical considerations in practical applications.

A multi-level fusion-based framework for multimodal fake news classification using semantic feature extraction

Abstract In the era of rapid digital news development, fake news poses a severe threat to society’s determination and authenticity, especially with the advent of online media platforms that facilitate the creation and dissemination of fabricated information. Although various techniques have been developed to discriminate between authentic and fake news, a practical fake news classification framework is still needed to automatically deliver high classification performance and impede the spread of misinformation. To fill this gap, this study proposes a multi-level fusion-based CNN with dual-conv layers-RNN (CDLR) framework that fuses Convolutional Neural Networks (CNN) with dual Conv layers, Recurrent Neural Networks (RNN), and classification module for multi-model fake news classification. The proposed framework fuses CNN (with dual-Conv layers) and RNN to enhance classification abilities and extract high-quality semantic textual and visual features to identify misinformation effectively. After pre-processing, the extracted weight matrix was fed to CNN (with dual Conv layers) to learn and extract deep visual features and an RNN for high-quality feature extraction from textual data or news articles for classification. Likewise, we designed a fusion mechanism to cross-validate the execution of our framework by considering different variants such as mean fusion, weighted-mean fusion, maximum fusion, and sum fusion. Finally, a classification module with a polynomial kernel was employed to categorize the extracted data as fake or real, for final classification. A comprehensive experiment analysis was carried out to evaluate the proposed framework's effectiveness by combining early and late fusion mechanisms with baseline methods on five extensive, fair, and diverse datasets. The accuracy of our framework was found to be 0.9725 on ISOT, 0.9107 on Fake vs. Real News, 0.9816 on WELFake, 0.5403 on FA-KES, and 0.9163 on Twitter datasets, indicating its robustness in classifying fake news compared to benchmark methods. Lastly, the study proposes some recommendations to mitigate the adverse effects of fake news based on the predictions made using our fusion-based framework.

Multi-Head Hierarchical Attention Framework with Multi-Level Learning Optimization Strategy for Legal Text Recognition

Owing to the rapid increase in the amount of legal text data and the increasing demand for intelligent processing, multi-label legal text recognition is becoming increasingly important in practical applications such as legal information retrieval and case classification. However, traditional methods have limitations in handling the complex semantics and multi-label characteristics of legal texts, making it difficult to accurately extract feature and effective category information. Therefore, this study proposes a novel multi-head hierarchical attention framework suitable for multi-label legal text recognition tasks. This framework comprises a feature extraction module and a hierarchical module. The former extracts multi-level semantic representations of text, while the latter obtains multi-label category information. In addition, this study proposes a novel hierarchical learning optimization strategy that balances the learning needs of multi-level semantic representation and multi-label category information through data preprocessing, loss calculation, and weight updating, effectively accelerating the convergence speed of framework training. We conducted comparative experiments on the legal domain dataset CAIL2021 and the general multi-label recognition datasets AAPD and Web of Science (WOS). The results indicate that the method proposed in this study is significantly superior to mainstream methods in legal and general scenarios, demonstrating excellent performance. The study findings are expected to be widely applied in the field of intelligent processing of legal information, improving the accuracy of intelligent classification of judicial cases and further promoting the digitalization and intelligence process of the legal industry.

Chinese pop culture’s development in the United States among Generation Z compared to South Korean pop culture

South Korean pop culture (K-pop) has achieved phenomenal business and cultural success in the United States, as evidenced by numerous hit songs, large-scale concerts, and record-breaking streaming and television viewership, particularly among Generation Z (Gen Z). In contrast, Chinese pop culture (C-pop) has experienced comparatively slower growth in the same market. While existing literature highlights the technological advancements and the business models of South Korean management companies as key drivers of K-pop’s success, there remains a research gap regarding the development and market performance of C-pop in the U.S. This qualitative study aimed to address that gap by conducting a comparative analysis of K-pop and C-pop in the U.S., specifically targeting Gen Z consumers. Utilizing a multiple case study methodology and a content analysis approach, this study examined the business models through numerical, textual, visual, and conceptual data derived from secondary sources. This study fills the gap in the existing literature by identifying the challenges facing C-pop in the U.S. market and proposing strategies for its growth. The findings suggest that a shift in creative direction, along with enhanced marketing strategies tailored to the U.S. market, is essential. Specifically, the study found that the American Gen Z audience prefers romance-themed and self-empowering content—whether in music or television—that establishes a personal connection and a sense of community. Therefore, it is strongly recommended that all aspects resonate with Gen Z population, a group of young individuals who experience loneliness.

MDFormer: Transformer-Based Multimodal Fusion for Robust Chest Disease Diagnosis

With the increasing richness of medical images and clinical data, abundant data support is provided for multimodal chest disease diagnosis methods. However, traditional multimodal fusion methods are often relatively simple, leading to insufficient exploitation of crossmodal complementary advantages. At the same time, existing multimodal chest disease diagnosis methods usually focus on two modalities, and their scalability is poor when extended to three or more modalities. Moreover, in practical clinical scenarios, missing modality problems often arise due to equipment limitations or incomplete data acquisition. To address these issues, this paper proposes a novel multimodal chest disease classification model, MDFormer. This model designs a crossmodal attention fusion mechanism, MFAttention, and combines it with the Transformer architecture to construct a multimodal fusion module, MFTrans, which effectively integrates medical imaging, clinical text, and vital signs data. When extended to multiple modalities, MFTrans significantly reduces model parameters. At the same time, this paper also proposes a two-stage masked enhancement classification and contrastive learning training framework, MECCL, which significantly improves the model’s robustness and transferability. Experimental results show that MDFormer achieves a classification precision of 0.8 on the MIMIC dataset, and when 50% of the modality data are missing, the AUC can reach 85% of that of the complete data, outperforming models that did not use two-stage training.

Research on multimodal English teaching model based on genetic algorithm

In recent years, English language teaching has undergone significant transformation due to the integration of technology and innovative teaching methodologies. Among the most promising advancements is the formation of multimodal teaching, which incorporates various forms of media. Despite the potential benefits associated with multimodal teaching, challenges remain in tailoring content to individual learning needs and optimizing the learning process for maximum effectiveness. To address these challenges, this research proposes a Multimodal English Teaching Framework (METF) to enhance the effectiveness and personalization of language learning. The framework integrates multiple teaching modalities, including text, audio, video, and interactive elements, to accommodate diverse learning preferences. It incorporates student interaction data (responses, time spent on tasks, and engagement levels), learner demographics, learning styles, and feedback records, thereby capturing the effectiveness of personalized content delivery. Pre-processing steps, including tokenization for text data, noise reduction for audio data, and normalization for resolution and interaction data, ensure data quality and consistency. Feature extraction techniques such as Convolutional Neural Networks (CNNs) for audio and video data and Term Frequency-Inverse Document Frequency (TF-IDF) for text data are applied to capture meaningful patterns. To enhance the fusion of multimodal data, feature-level fusion is implemented, integrating recovered characteristics from multiple modalities into a unified representation that supports more accurate decision-making. This research establishes the Resilient Genetic Algorithm (RGA) to enhance the adaptive learning process by ensuring robustness and avoiding premature convergence. Experimental results demonstrate the potential of this multimodal approach, achieving 94.2% accuracy, a convergence time of 45%, task engagement of 89.4%, and personalization effectiveness of 91.7%. These results indicate a high value of accuracy in multimodal English teaching. This research highlights the applicability of algorithms in optimizing complex, multimodal educational environments.

Smartphone-based text obtained via passive sensing as it relates to direct suicide risk assessment.

Recent research highlights the dynamics of suicide risk, resulting in a shift toward real-time methodologies, such as ecological momentary assessment (EMA), to improve suicide risk identification. However, EMA's reliance on active self-reporting introduces challenges, including participant burden and reduced response rates during crises. This study explores the potential of Screenomics-a passive digital phenotyping method that captures intensive, real-time smartphone screenshots-to detect suicide risk through text-based analysis. Seventy-nine participants with past-month suicidal ideation or behavior completed daily EMA prompts and provided smartphone data over 28days, resulting in approximately 7.5 million screenshots. Text from screenshots was analyzed using a validated dictionary encompassing suicide-related and general risk language. Results indicated significant associations between passive and active suicidal ideation and suicide planning with specific language patterns. Detection of words related to suicidal thoughts and general risk-related words strongly correlated with self-reported suicide risk, with distinct between- and within-person effects highlighting the dynamic nature of suicide risk factors. This study demonstrates the feasibility of leveraging smartphone text data for real-time suicide risk detection, offering a scalable, low-burden alternative to traditional methods. Findings suggest that dynamic, individualized monitoring via passive data collection could enhance suicide prevention efforts by enabling timely, tailored interventions. Future research should refine language models and explore diverse populations to extend the generalizability of this innovative approach.

An Integrated Artificial Intelligence Model for Understanding User Intent: Leveraging Internet Text Data Mining and Intelligence Influence Analysis from Online Reviews and Social Media

An Integrated Artificial Intelligence Model for Understanding User Intent: Leveraging Internet Text Data Mining and Intelligence Influence Analysis from Online Reviews and Social Media

Emotion Recognition System Using Machine Learning

This paper introduces the concept of an "Emotion Recognition system using machine learning (ERS)" designed for recognizing and analysis of the emotions. The emotions that are detected are through speech, text or facial based data. It addresses key challenges, such as analyzing the patient’s emotions through the facial expressions and the speech to monitor mental health, provide accurate feedback based on the text inputs and speech while using customer care feedbacks. The Emotion Recognition system using machine learning (ERS) uses different models for each data to provide accurate results. For the facial expressions sequential model is used, for speech Convolutional Neural Network (CNN) and Recurrent Neural Network(RNN) are used. The text data uses transformer-based models.

Understanding Public Opinion through Sentiment Analysis: A Case Study Using the Sentiment140 Dataset

The rise of social media has provided an unprecedented opportunity to analyze public opinion at scale. Sentiment analysis, a key technique in natural language processing (NLP), enables the automatic classification of opinions expressed in textual data. This study explores the effectiveness of sentiment analysis using the Sentiment140 dataset, a collection of tweets labeled as positive, negative, or neutral. We compare machine learning and deep learning models, including logistic regression, support vector machines (SVM), convolutional neural networks (CNN), and long short-term memory (LSTM) networks. Our results highlight the comparative strengths of each model in classifying sentiments with varying degrees of accuracy. This research contributes to the ongoing development of AI-powered sentiment analysis by evaluating model performance on real-world social media data.

Automatic theme and motif identification in large-scale English literary corpora using deep learning approaches

The identification of themes and motifs in literary texts is a fundamental aspect of literary analysis, traditionally performed through manual annotation and expert interpretation. However, the increasing availability of large-scale English literary corpora presents new challenges and opportunities for automated analysis. This paper proposes a deep learning (DL)-based framework for automatically detecting themes and motifs in extensive literary collections. The dataset comprises diverse sources, including classic literature, modern fiction, and poetry, ensuring a broad representation of thematic structures. A rigorous preprocessing pipeline is applied, involving stop word removal and tokenization to refine textual data. For feature extraction, Word2Vec is utilized to capture semantic relationships between words. The core novelty of this research lies in the implementation of a Duelist Algorithm-optimized Bi-directional Long Short-Term Memory (DAO-BiLSTM) model, which enhances the model’s ability to detect and classify recurring thematic elements with high accuracy. The proposed method achieves an accuracy of 96.24%, recall of 97.32%, precision of 95.6%, and an F1-score of 94.7%, demonstrating superior performance over existing methods. The model is implemented in Python 3.9 using TensorFlow in a high-performance computing environment, ensuring efficient processing of large-scale textual data. Experimental results illustrate the effectiveness of the proposed approach in identifying complex motifs and themes across various literary genres. These findings highlight the potential of DL in augmenting literary analysis, enabling large-scale, data-driven thematic exploration that complements traditional human-driven methodologies.

Clinical insights: A comprehensive review of language models in medicine.

This paper explores the advancements and applications of language models in healthcare, focusing on their clinical use cases. It examines the evolution from early encoder-based systems requiring extensive fine-tuning to state-of-the-art large language and multimodal models capable of integrating text and visual data through in-context learning. The analysis emphasizes locally deployable models, which enhance data privacy and operational autonomy, and their applications in tasks such as text generation, classification, information extraction, and conversational systems. The paper also highlights a structured organization of tasks and a tiered ethical approach, providing a valuable resource for researchers and practitioners, while discussing key challenges related to ethics, evaluation, and implementation.