Topic Modeling Approach Research Articles

Developing an Integrated Topic Modeling Approach for Text-Based Knowledge Accumulation

Developing an Integrated Topic Modeling Approach for Text-Based Knowledge Accumulation

Information Security Research in the Information Systems Discipline: A Thematic Review and Future Research Directions

Information security continues to grow in importance in all aspects of society and therefore evolves as a prevalent research area. The information systems (IS) discipline offers a unique perspective to move this stream of literature forward. Using a semiautomated thematic analysis approach based on the topic modeling technique, we review a broad range of information security literature to investigate how we may integrate and advance our understanding of information security. Our analysis reveals four major themes, including information security policy (ISP) compliance, motivations and susceptibility, software security decisions, and firm security strategy. We also identify a theme of security in broader contexts, in which studies consider the societal impacts of information security and online hacker behavior. This review contributes to IS literature by 1) synthesizing the broad range of information security research published in the top journals of the field, moving beyond prior reviews focusing on a narrower scope such as individual ISP compliance; 2) identifying major themes and proposing an overarching research framework encompassing these themes and their interconnections, allowing us to envision future research directions; and 3) enumerating a semiautomated topic modeling approach that other researchers can employ.

Exploring the technological leaders using tire industry patents: A topic modeling approach

Researchers have long been interested in the transportation industry, recognizing its significant impact on human societies. Specifically, the automobile industry and its components have garnered special attention due to the evolving nature of transportation technologies. With this in mind, this study aims to explore the current and future technologies within the tire industry as a technological-based component of the automotive industry. Additionally, the research examines the key topics, leaders of technology, and strategies of technology-owning companies in these topics. To carry out this investigation, we employed a combination of methods including patent analysis and topic modeling. The selection of leading technology companies was based not only on the number of patents but also on the impact and quality of their patents, which was carefully evaluated. By leveraging these methods and results, tire companies and other industries can gain a competitive edge in technology, acquire insights into the current and future market share of leading companies, and develop effective roadmaps for their own advancement.

Philosophical research in Brazil: A structural topic modeling approach with a focus on temporal and gender trends

AbstractThis paper employs structural topic modeling (STM) to describe the academic philosophy landscape in Brazil. Based on a public national database, a corpus consisting of 12,515 abstracts of monographs defended in philosophy graduate programs between 1991 and 2021 was compiled. The final STM model identified 74 meaningful research topics, clustered into 7 thematic categories. This study discusses the prevalence of the most significant topics and categories, their trends across three decades, and their (positive or negative) association with the supervisor's gender. Results show the first empirical evidence that Brazilian philosophical research exhibits a greater focus on philosophers than on specific themes or problems. Moreover, by visualizing the variations in topic prevalence over time, it was possible to track the rise or decline of the major interest categories and topics. Finally, results also show which topics are more influenced or less influenced by gender.

Talking past each other: Institutional complexity in public projects

Large-scale public projects often experience delays and cost overruns, in part because they require extensive coordination between numerous organizations, both private and public. In this paper, we draw on the perspective of new institutionalism to explore differences between stakeholders in large-scale projects. We develop a topic modeling approach based on a state-of-the-art computational text analysis technique and apply it to a set of transcripts from parliamentary select committee hearings in the United Kingdom to uncover cultural-cognitive differences that reflect a variety of institutional logics. The application of this technique to four different projects reveals significant institutional differences between groups of actors. We argue that greater awareness of these institutional differences may support attempts to understand why large-scale projects experience such significant management challenges and how such challenges could be addressed. We use data for the period 2000–2019 from two different sectors, defence, and public transport.

Simplifying synthesis of the expanding glioblastoma literature: a topic modeling approach.

Our study aims to discover the leading topics within glioblastoma (GB) research, and to examine if these topics have "hot" or "cold" trends. Additionally, we aim to showcase the potential of natural language processing (NLP) in facilitating research syntheses, offering an efficient strategy to dissect the landscape of academic literature in the realm of GB research. The Scopus database was queried using "glioblastoma" as the search term, in the "TITLE" and "KEY" fields. BERTopic, an NLP-based topic modeling (TM) method, was used for probabilistic TM. We specified a minimum topic size of 300 documents and 5% probability cutoff for outlier detection. We labeled topics based on keywords and representative documents and visualized them with word clouds. Linear regression models were utilized to identify "hot" and "cold" topic trends per decade. Our TM analysis categorized 43,329 articles into 15 distinct topics. The most common topics were Genomics, Survival, Drug Delivery, and Imaging, while the least common topics were Surgical Resection, MGMT Methylation, and Exosomes. The hottest topics over the 2020s were Viruses and Oncolytic Therapy, Anticancer Compounds, and Exosomes, while the cold topics were Surgical Resection, Angiogenesis, and Tumor Metabolism. Our NLP methodology provided an extensive analysis of GB literature, revealing valuable insights about historical and contemporary patterns difficult to discern with traditional techniques. The outcomes offer guidance for research directions, policy, and identifying emerging trends. Our approach could be applied across research disciplines to summarize and examine scholarly literature, guiding future exploration.

Evolving role of public-private stakeholders in technology mediated teaching

Purpose Post-pandemic sovereign authorities in several economies have nudged primary education institutions to adopt platform-based teaching. The shift to platform technology attempts to ensure continuity in the teaching–learning process. In the context of predominantly digitally mediated teaching process, this shift may exacerbate disparities and social injustice by limiting access to primary education in resource-constrained developing economies. The purpose of this study is to explore the efficacy of such a digital framework provided by government and private partners and the challenges faced by the teachers in absence of proper scaffolding. Design/methodology/approach Using an integrative theoretical framework that is composed of capability theory, technology adoption theories and the scaffolding framework, this paper analyses the challenges faced by primary school teachers when adapting to platform-based teaching. Social media analytics along with text analytics using Natural Language Processing and latent Dirichlet allocation-based topic modelling approach to extract latent topics or themes used by users during their tweets related to e-teaching. Findings The findings of this study highlight that adopting a platform-based and hybrid approach improves access to education and flexibility and highlights the importance of scaffolds in achieving desired learning outcomes. EdTech companies can play a significant role through private-public partnership models to offer technical scaffold. Collaborative efforts between educational institutions and EdTech service providers are essential for ensuring the long-term sustainability of platform-based teaching and learning. Originality/value After the pandemic, there has been no published literature available which examined the role of scaffolds and EdTech companies in ensuring digital ecosystem for better teaching–learning outcome through platforms.

Technology roadmapping for the e-commerce sector: A text-mining approach

Due to rapid technological advancements and shifting consumer dynamics, the e-commerce industry has undergone rapid transformation. Further, emerging technologies, including artificial intelligence, blockchain, and augmented reality, have the ability to revolutionize the e-commerce landscape. However, limited studies have mapped the technological evolution in the e-commerce sector. Therefore, this study offers a multi-faceted forecast of e-commerce technologies by combining patent analysis, semantic network analysis, and topic modelling approaches. The study analyzes 4113 patents and categorizes them into two time periods: 2001–2012 and 2013–2023. This study constructs co-word maps, defines the relationships between the words, and examines how the co-word relationships have changed over the time period. In the period between 2001 and 2012, the patents are categorized into four clusters: information processing and communication technologies, product purchase and sale management, e-commerce systems, and digital transaction and security management. For the period spanning between 2013 and 2023, the patents are classified into seven clusters: user experience, integrating devices, product information management, e-commerce infrastructure, advanced e-commerce authentication, pricing and payment solution, platform development, and e-commerce optimization. It aims to bridge theory and practice by developing a roadmap of near-, mid-, and far-future e-commerce technologies. The study highlights implications for practice and theory.

Assessing Public Perceptions of Virtual Primary Care During the COVID-19 Pandemic in the UK, Germany, Sweden, and Italy: A Topic Modeling Approach

Assessing Public Perceptions of Virtual Primary Care During the COVID-19 Pandemic in the UK, Germany, Sweden, and Italy: A Topic Modeling Approach

Analyzing Digital Propaganda on X through Topic Modelling Approach: The Case of Afghanistan

The dissemination of propaganda has become a formidable weapon, with nation-states exploiting social media platforms to engineer narratives favorable to their geopolitical interests. This study delved into Indo-Pakistan states' orchestrated propaganda in the wake of the Taliban's recapture of Afghanistan. This period is marked by a sophisticated blend of propaganda strategies to mold online discourse. Utilizing a dataset derived from X (formerly Twitter), the research examines how Pakistan and India leveraged information warfare to advance their politico-diplomatic narratives, shedding light on the propaganda strategies of such digital warfare and the emergence of narratives and counter-narratives. The findings highlight the instrumental role of propaganda strategies in amplifying political, diplomatic, security, and humanitarian narratives and manipulating public discourse, with distinct tactics and officially approved state maneuvers identified through the topic modeling and analyzing key strategies of propaganda.

A Comparative Study of Topic Modelling Approaches for User-generated Point of Interest Data

This study aims to enhance urban planning and management by harnessing the power of machine learning (ML) andbig data. We focus on Urban Functional Zones (UFZs), the fundamental units for human socio-economic activities. Ourmethodology involves compiling Point of Interest (POI) data from various sources for comprehensive analysis. Weemploy various topic modeling approaches such as Latent Dirichlet Allocation (LDA), Latent Semantic Index (LSI),Hierarchical Dirichlet Process (HDP), and Top2Vec. Our principal results reveal significant differences in theperformance and coherence of these models on short text documents. Consequently, our major conclusion isidentifying the better-performing topic model for classifying UFZs from POI data. We also explore four textpreprocessing steps to optimize the performance of the topic models. This study contributes to the field by providinga nuanced understanding of UFZs, paving the way for future data-driven urban planning and management.

Sentiment Analysis on 'HelloTalk' App Reviews Using NRC Emotion Lexicon and GoEmotions Dataset

During the post-pandemic period, the interest in foreign language learning surged, leading to increased usage of language-learning apps. With the rising demand for these apps, analyzing app reviews becomes essential, as they provide valuable insights into user experiences and suggestions for improvement. This research focuses on extracting insights into users' opinions, sentiments, and overall satisfaction from reviews of HelloTalk, one of the most renowned language-learning apps. We employed topic modeling and emotion analysis approaches to analyze reviews collected from the Google Play Store. Several experiments were conducted to evaluate the performance of sentiment classification models with different settings. In addition, we identified dominant emotions and topics within the app reviews using feature importance analysis. The experimental results show that the Random Forest model with topics and emotions outperforms other approaches in accuracy, recall, and F1 score. The findings reveal that topics emphasizing language learning and community interactions, as well as the use of language learning tools and the learning experience, are prominent. Moreover, the emotions of 'admiration' and 'annoyance' emerge as significant factors across all models. This research highlights that incorporating emotion scores into the model and utilizing a broader range of emotion labels enhances model performance.

Public Perceptions of EV Charging Infrastructure: A Combined Sentiment Analysis and Topic Modeling Approach

Public Perceptions of EV Charging Infrastructure: A Combined Sentiment Analysis and Topic Modeling Approach

A text mining approach to explore factors influencing consumer intention to use metaverse platform services: Insights from online customer reviews

Metaverse technology has recreated real-world surroundings into a virtual setting that has caught the attention of many consumers and brands. Metaverse platforms are facilitating brands to sell physical and digital versions of their products, and consumers are eagerly purchasing these products. Owing to this dramatic shift, it is crucial to understand the factors impacting user adoption of metaverse platform services. The present study leverages an LDA-based topic modeling approach to analyze 17136 online user reviews of the Highrise Virtual Metaverse mobile platform. The study revealed ten factors impacting consumer intention to use the metaverse platform services that are asset buy-sell experience, availability of virtual assets, visually appealing virtual environment, enjoyment from virtual events, social interaction, social activity, positive experience with the metaverse ecosystem, mobile app-related issues, user account-related issues, and the feature of watch ad-to earn. Building upon the uses and gratification theory and flow theory, the study explained that consumers gain three major gratifications and experience a flow state while using metaverse platform services. Furthermore, it explores the emotions expressed by users within the metaverse platform. Correspondence analysis is performed to understand the factors leading to positive and negative user experiences with the platform services. The study presents crucial practical implications for brand marketers and metaverse platform developers.

Selecting textual analysis tools to classify sustainability information in corporate reporting

Information on firms' sustainability often partly resides in unstructured data published, for instance, in annual reports, news, and transcripts of earnings calls. In recent years, researchers and practitioners have started to extract information from these data sources using a broad range of natural language processing (NLP) methods. While there is much to be gained from these endeavors, studies that employ these methods rarely reflect upon the validity and quality of the chosen method—that is, how adequately NLP captures the sustainability information from text. This practice is problematic, as different NLP techniques lead to different results regarding the extraction of information. Hence, the choice of method may affect the outcome of the application and thus the inferences that users draw from their results. In this study, we examine how different types of NLP methods influence the validity and quality of extracted information. In particular, we compare four primary methods, namely (1) dictionary-based techniques, (2) topic modeling approaches, (3) word embeddings, and (4) large language models such as BERT and ChatGPT, and evaluate them on 75,000 manually labeled sentences from 10-K annual reports that serve as the ground truth. Our results show that dictionaries have a large variation in quality, topic models outperform other approaches that do not rely on large language models, and large language models show the strongest performance. In large language models, individual fine-tuning remains crucial. One-shot approaches (i.e., ChatGPT) have lately surpassed earlier approaches when using well-designed prompts and the most recent models.

Exploring the surveillance technology discourse: a bibliometric analysis and topic modeling approach.

The prevention of crime is a multifaceted challenge with legal, political, and cultural implications. Surveillance technologies play a crucial role in assisting law enforcement and other relevant parties in this mission. Drones, cameras, and wiretaps are examples of such devices. As their use increases, it becomes essential to address related challenges involving various stakeholders and consider cultural, political, and legal aspects. The objective of this study was to analyze the impact of surveillance technologies and identify commonalities and differences in perspectives among social media users and researchers. Data extraction was performed from two platforms: Scopus (for academic research papers) and platform X (formerly known as Twitter). The dataset included 88,989 tweets and 4,874 research papers. Topic modeling, an unsupervised machine learning approach, was applied to analyze the content. The research results revealed that privacy received little attention across the datasets, indicating its relatively low prominence. The military applications and their usage have been documented in academic research articles as well as tweets. Based on the empirical evidence, it seems that contemporary surveillance technology may be accurately described as possessing a bi-directional nature, including both sousveillance and surveillance, which aligns with Deleuzian ideas on the Panopticon. The study's findings also indicate that there was a greater level of interest in actual applications of surveillance technologies as opposed to more abstract concepts like ethics and privacy.

What Do Flutter Developers Ask About? An Empirical Study on Stack Overflow Posts

Since Google launched Flutter, an open-source framework, in 2017, many companies and software developers have turned to its use owing to its cross-platform feature. Other attractive features include hot reloading, a rich widget library, and improved performance compared to other cross-platform frameworks. Despite the rise in the use and adoption of the framework, little has been done to understand developers’ challenges. This study aims to understand what Flutter developers post on Stack Overflow – a popular Q&A website for developers. Analyzing such posts would help us understand the challenges faced by Flutter developers. To meet this goal, the study used a topic modeling approach to analyze all “flutter” tagged posts between 2019 and 2023. This study revealed state management, widgets, navigation, packages, and persistence as some areas developers face challenges. Further, the study established that there is a growth in the number of Flutter-related posts and developers. While the Flutter framework is promising for companies and software developers, this study points out areas where Flutter trainers and developers should emphasize. Flutter Framework developers should provide more documentation and support as the language matures.

A decadal study on identifying latent topics and research trends in open access LIS journals using topic modeling approach

A decadal study on identifying latent topics and research trends in open access LIS journals using topic modeling approach

The information content of rating action reports: A topic modeling approach

This paper examines the information content in Moody's rating action reports, which Moody's releases concurrently with its rating actions. Using a topic modeling approach, we identify two informative topics after controlling for rating characteristics and report tones. We find that unfavorable discussion of the Metrics topic (about issuers' financial performance) generates a significant positive market reaction, while unfavorable discussion of the Liquidity topic (about issuers' liquidity profile) generates a significant negative market reaction around rating actions. Together, these two topics increase the explanatory power of market reaction around rating actions by over 10%. Further analysis reveals significant associations between these two topics and future default likelihood and debt-paying ability, and the associations are consistent with market reaction to these topics. Although investors have less confidence in rating reports following the Dodd-Frank Act, these two topics are more predictive of future default probabilities. Overall, we find that rating reports topics provide important information to investors as they significantly predict future default likelihood and financial performance.

Analysis of Current Research in the Field of Sustainable Employment Based on Latent Dirichlet Allocation

In the current context of sluggish global economic recovery, widening regional disparities, and little room for policy error, the global employment situation is unprecedentedly severe. Therefore, research on sustainable employment is critical. In this study, we utilized a Latent Dirichlet Allocation (LDA) topic modeling approach to identify four central topics within the realm of sustainable employment: “Career Sustainability and Employee Development”, “Socio-Economic Dynamics and Strategies for Sustainable Employment”, “Study on Age, Health, and Sustainable Workforce in the Labor Market”, and “Research on the Impact of Workplace Interventions and Psychological Health Support”. The findings revealed the multidimensional, complex nature of sustainable employment issues. This study expands the theoretical horizon of sustainable employment and proposes multidimensional strategic suggestions for practical applications, including regularly assessing and updating organizations’ leadership styles and employee development strategies, offering customized career development programs and skill training, and formulating inclusive employment policies and supportive workplace environments in order to enhance employee job satisfaction and overall well-being, thereby effectively improving the practical outcomes of sustainable employment and bringing about long-term stability and positive impacts on the labor market. These results provide essential theoretical and practical contributions for policymakers, organizational leaders, and anyone seeking sustainable employment.