Latent Dirichlet Allocation Method Research Articles

Application of convolutional neural networks in medical images: a bibliometric analysis.

In the field of medical imaging, the rapid rise of convolutional neural networks (CNNs) has presented significant opportunities for conserving healthcare resources. However, with the wide spread application of CNNs, several challenges have emerged, such as enormous data annotation costs, difficulties in ensuring user privacy and security, weak model interpretability, and the consumption of substantial computational resources. The fundamental challenge lies in optimizing and seamlessly integrating CNN technology to enhance the precision and efficiency of medical diagnosis. This study sought to provide a comprehensive bibliometric overview of current research on the application of CNNs in medical imaging. Initially, bibliometric methods were used to calculate the frequency statistics, and perform the cluster analysis and the co-citation analysis of countries, institutions, authors, keywords, and references. Subsequently, the latent Dirichlet allocation (LDA) method was employed for the topic modeling of the literature. Next, an in-depth analysis of the topics was conducted, and the topics in the medical field, technical aspects, and trends in topic evolution were summarized. Finally, by integrating the bibliometrics and LDA results, the developmental trajectory, milestones, and future directions in this field were outlined. A data set containing 6,310 articles in this field published from January 2013 to December 2023 was complied. With a total of 55,538 articles, the United States led in terms of the citation count, while in terms of the publication volume, China led with 2,385 articles. Harvard University emerged as the most influential institution, boasting an average of 69.92 citations per article. Within the realm of CNNs, residual neural network (ResNet) and U-Net stood out, receiving 1,602 and 1,419 citations, respectively, which highlights the significant attention these models have received. The impact of coronavirus disease 2019 (COVID-19) was unmistakable, as reflected by the publication of 597 articles, making it a focal point of research. Additionally, among various disease topics, with 290 articles, brain-related research was the most prevalent. Computed tomography (CT) imaging dominated the research landscape, representing 73% of the 30 different topics. Over the past 11 years, CNN-related research in medical imaging has grown exponentially. The findings of the present study provide insights into the field's status and research hotspots. In addition, this article meticulously chronicled the development of CNNs and highlighted key milestones, starting with LeNet in 1989, followed by a challenging 20-year exploration period, and culminating in the breakthrough moment with AlexNet in 2012. Finally, this article explored recent advancements in CNN technology, including semi-supervised learning, efficient learning, trustworthy artificial intelligence (AI), and federated learning methods, and also addressed challenges related to data annotation costs, diagnostic efficiency, model performance, and data privacy.

Insights on emotional labour research: a topic modelling approach

Purpose Emotional labour (EL) is a complex phenomenon that has received increasing attention in recent years due to its impact on employee’s well-being and job satisfaction. For a comprehensive understanding of the evolving field of EL, it is important to extract different research trends, new developments and research directions in this domain. The study aims to reveal 13 prominent research topics based on the topic modelling analysis. Design/methodology/approach Using latent Dirichlet allocation (LDA) method, topic modelling is done on 1,462 journal research papers published between 1999 and 2023, extracted from the Scopus database using the keyword “EL”. Findings The analysis identifies several emerging trends in EL research, including emotional regulation training and job redesign. Similarly, the topics like EL strategies, cultural differences and EL, EL in hospitality, organizational support and EL, EL and gender and psychological well-being of nursing workers are popular research topics in this domain. Research limitations/implications The findings provide valuable insights into the current state of EL research and can provide a direction for future research as well as assist organizations to design practices aimed at improving working conditions for employees in various industries. Originality/value Topic modelling on emotional labor is done. The paper identifies specific topics or clusters related to emotional labor, quantifies these topics using topic modeling, adds empirical rigor, and allows for comparisons across different contexts.

Performance evaluation of Latent Dirichlet Allocation on legal documents

Latent Dirichlet Allocation (LDA) is an algorithm with the capability of processing large amount of text data. In this study, the LDA is used to produce topic modelling of topic clusters from corpus of legal texts generated under 4 topics within Nigeria context Employment Contract, Election Petition, Deeds, and Articles of Incorporation. Each topic has a substantial number of articles and the LDA method proves effective in extracting topics and generating index words that are in each topic cluster. At the end of experimentation, results are compared with manually pre-annotated dataset for validation purpose and the results show high accuracy. The LDA output shows optimal performance in the word indexing processing for Election Petition as all the documents annotated under the topic were accurately classified.

Real-Time Twitter Trends Analysis Using Latent Dirichlet Allocation and Machine Learning

As far as social media is concerned twitter has become a major source of public data in the form of tweets. Twitter is an emerging source of large textual data(big data).People can easily express their opinions , reviews, interests and tastes about a particular event, topic ,product etc., occurring worldwide. This makes twitter a good source of valuable data which can be further used to perform sentiment analysis, know trending topics of public intrests and public opinion on a particular product or event which can be beneficial for business growth and political parties to know public choices and they can take actions accordingly. The extraction of meaningful insights from the vast and dynamic stream of social media data is greatly facilitated by the implementation of topic modeling and sentiment analysis on Twitter tweets. The identification of recurring themes or subjects within a collection of tweets, which is the essence of topic modeling, serves to reveal patterns in discussions, thereby aiding in the comprehension of prevalent topics and trends. Concurrently, sentiment analysis, through the evaluation of the emotional tone of tweets, enables the discernment of whether the expressed sentiments are positive, negative, or neutral. The combination of these techniques provides researchers and businesses with valuable perspectives on public opinions, emerging issues, and user sentiments, thereby empowering them to make informed decisions and develop effective engagement strategies in the ever-evolving landscape of social media. But twitter data(tweets) are unstructured ,contains noisy data, urls , stop words, re-tweets video ,emoji etc., which need to be cleaned and preprocessed to perform proper sentiment analysis and use it to extract meaningful effective insights from it. This paper focuses on various methods of topic modeling and discovering latent topics ,text-mining approaches ,micro-blogging methods used in various researches .This paper focuses on latent dirichlet allocation method of topic modeling and text-mining to discover latent topics in tweets, micro-blogging ,text-mining approaches .A proper survey is done on previous researches on this topic in this article. Keywords:-LDA, text-mining, topic modeling, micro-blogging ,machine learning ,sentiment analysis, NLP.

Predicting likelihood of fraud among financial distressed firms in Malaysia using textual analysis

This research paper aims to analyze and predict fraud patterns among failed companies in Malaysia. The approach involves utilizing textual analysis on the management discussion and analysis (MD&A) section within the annual reports. The dataset is subjected to text clustering to group companies based on similar financial characteristics. This clustering process entails several steps, including data conversion, collation, and summarization into a structured format, followed by text pre-processing to cleanse the dataset. Notably, RapidMiner Studio software was utilized to extract data for the study. Subsequently, the documents are clustered using both the K-means and latent dirichlet allocation (LDA) methods. Upon examining a sample of 22 failed companies in the year 2020, the study reveals that financially distressed companies exhibit prominent financial negativity and utilize litigious financial terms within their MD&A sections. These linguistic traits are found to be closely associated with seven distinct characteristics of fraudulent firms. This preliminary findings provide compelling evidence that financial pressure may serve as a triggering factor for fraudulent activities within companies.

Characteristics of students’ learning behavior preferences — an analysis of self-commentary data based on the LDA model

How to better grasp students’ learning preferences in the environment of rapid development of engineering and science and technology so as to guide them to high-quality learning is one of the important research topics in the field of educational technology research today. In order to achieve this goal, this paper utilizes the LDA (Latent Dirichlet Allocation) model for text mining of the survey results on the basis of a survey on students’ self-perception evaluation. The results show that the LDA model is capable of extracting terms from text, fuzzy identifying groups of students at different levels and presenting potential logical relationships between the groups, and further analyzing the learning preferences of students at different levels for IT courses. Based on the student’s learning needs, this paper proposes recommendations for developing students’ learning effectiveness. The LDA method proposed in this paper is a feasible and effective method for assessing students’ learning dynamics as it generates cognitive content about students’ learning and allows for the timely discovery of students’ learning expectations and cutting-edge dynamics.

Navigating Success in Carbon Offset Projects: A Deep Dive into the Determinants Using Topic Modeling

Carbon offset projects play a crucial role in tackling the global challenge of climate change. However, there is limited understanding of the factors contributing to the success of a carbon offset project. In this study, we utilize the latent Dirichlet allocation method to extract topics from the descriptions of carbon offset projects sourced from the Gold Standard Foundation. Our findings reveal that projects encompassing both safety and efficient energy solutions for households command higher prices. These results imply that an effective carbon offset project should mitigate individual household emissions while enhancing safety. Our research carries significant implications for stakeholders involved in carbon offset projects and can serve as a foundation for policy formulation and standard regulations.

The changing role of hotel attributes in destination competitiveness throughout a crisis

Purpose By conducting a longitudinal study, this study aims to investigate how the role of hotel attributes in destination competitiveness changed through the stages of pre-, amid and recovery from the crisis. Design/methodology/approach First, the latent Dirichlet allocation method was used to identify hotel attributes from 15,137 online reviews, and then a sentiment analysis was performed to determine tourist satisfaction with the subject attributes. Second, separate asymmetric impact competitor analyses were conducted for the three stages of the crisis, and their results were compared with understand how the role of the hotel attributes changed throughout the crisis. Findings The results revealed that the impacts of hotel attributes on tourist satisfaction and destination competitiveness differed significantly at each stage of the crisis. Research limitations/implications This research expands the existing literature by offering valuable insights by elucidating the changing characteristics of hotel attributes at each crisis stage. The results extend the body of knowledge in destination management by providing evidence on the validity of asymmetric impact competitor analysis. Originality/value To fully understand the impact of a crisis (e.g. COVID-19) on destination competitiveness with a focus on the hotel sector, this research conducted a longitudinal study that covers three stages of the crisis (i.e. pre-, amid and post-crisis). Moreover, unlike previous studies, this research considers the asymmetric relationships between service attributes and overall tourist satisfaction, as well as competitors’ information.

Latent Dirichlet Allocation (LDA) topic models for Space Syntax studies on spatial experience

Spatial experience has been extensively researched in various fields, with Space Syntax being one of the most widely used methodologies. Multiple Space Syntax techniques have been developed and used to quantitively examine the relationship between spatial configuration and human experience. However, due to the heterogeneity of syntactic measures and experiential issues in the built environment, a systematic review of socio-spatial topics has yet to be developed for Space Syntax research. In response to this knowledge gap, this article employs an ‘intelligent’ method to classify and systematically review topics in Space Syntax studies on spatial experience. Specifically, after identifying 66 articles using the ‘Preferred Reporting Items for Systematic reviews and Meta-Analyses’ (PRISMA) framework, this research develops generative probabilistic topic models to classify the articles using the Latent Dirichlet Allocation (LDA) method. As a result, this research automatically generates three architectural topics from the collected literature data (A1. Wayfinding behaviour, A2. Interactive accessibility, and A3. Healthcare design) and three urban topics (U1. Pedestrian movement, U2. Park accessibility, and U3. Cognitive city). Thereafter it qualitatively examines the implications of the data and its LDA classification. This article concludes with an examination of the limitations of both the methods and the results. Along with demonstrating a methodological innovation (combining PRISMA with LDA), this research identifies critical socio-spatial concepts and examines the complexity of Space Syntax applications. In this way, this research contributes to future Space Syntax research that empirically investigates the relationships between syntactic and experiential variables in architectural and urban spaces. The findings support a detailed discussion about research gaps in the literature and future research directions.

Product improvement in a big data environment: A novel method based on text mining and large group decision making

Product improvement has become a multifaceted and uncertain endeavour for manufacturers in an increasingly competitive business environment. Online platforms have emerged to solicit consumer opinions and product feedback. However, product improvement requires a critical yet complex decision-making approach for manufacturers. Managers face the challenge of identifying the most effective decision-making methodology for product improvement, especially in a big data environment. In this research, we comprehensively evaluate different product improvement decision-making methodologies through a series of experimental investigations. Specifically, three different experiments are conducted, including: i) an initial selection guided by intuitive perception, ii) expert decision-making, and iii) a hybrid method that incorporates consumer big data and large-scale group decision-making. Product criteria sets are categorized using the Latent Dirichlet Allocation (LDA) method, while the importance of these criteria is determined by applying the TextRank and Word2Vec algorithms. Our empirical results show that the mixed method, which utilizes text-mining techniques in conjunction with large-group decision-making, provides a more reliable and effective approach to facilitating product improvement.

Knowledge spillovers and entrepreneurial ecosystem: A systematic literature review

The purpose of this study was to synthesize the literature that has been established on the interrelationship of “knowledge spillovers and entrepreneurial ecosystem” through a systematic literature review. The query strings generated 85 articles, of which 51 were selected for final review. The analysis was performed using MySLR software that uses the Latent Dirichlet Allocation (LDA) method, and the effectiveness was measured through the perplexity score and coherence value of the topics. The results reveal that although a plethora of literature is available on these two topics individually, the mutual relationship between these two concepts is understudied. A stronger research interest was observed after 2018 which indicates the infancy of the topic as a whole, which is yet to be explored and empirically tested, specifically in developing economies. Most of the research was focused on the role of universities, innovation, geographic dynamics and policies with respect to entrepreneurial ecosystems. Results reveal the disparity of knowledge spillovers with regard to other contributors; significantly lesser attention was paid to its role in relation to the entrepreneurial ecosystem. Along with critical analysis and synthesis of literature, future directions have also been presented in the paper.

Dividend announcements and stock returns: a retrospective analysis

Purpose This study aims to provide insights into different aspects of the extant literature on the effects of dividend announcements. Along with other outputs of a bibliometric study, this study provides deeper insights into the concentration of the extant literature and suggest future research agendas. Design/methodology/approach This study uses the bibliometric, network and content analysis of the dividend announcement literature indexed in Scopus. This study presents the temporal analysis, the network of authors, countries, author citations and the co-occurrence of author keywords. This study provides the concentration of the extant literature in three clusters and unearth some key future research areas. This study uses the latent Dirichlet allocation method for robustness. Findings A total of 54 documents examining the US sample have received 1,804 citations. Interestingly, the first article on emerging markets was published in 2002, when at least 34 articles on developed markets had already been published from 1982 to 2001. The content analysis of top-cited literature unveils diverse insights into dividend announcements’ effects on financial markets. Contagion effects negatively impact non-announcing banks, particularly larger ones. Dividend maintenance affects stock market momentum, influencing loser returns. While current dividend/earnings news may not predict future company performance, information content dominates bond market reactions to post-dividend announcements. Concomitantly, while financially constrained firms exhibit short-term gains but worse long-term performance following dividend increases, larger stock dividends send stronger market signals in China. Originality/value This study significantly contributes to the bibliometric and content analysis literature by analyzing the sample documents based on the sample examined. To the best of the authors’ knowledge, no previous bibliometric study in this domain has been conducted to explore the markets (developed and emerging) to which the samples examined belong and the quality of publications from developed and emerging markets.

MPCM: Multi-modal User Portrait Classification Model Based on Collaborative Learning

A social-media user portrait is an important means of improving the quality of an Internet information service. Current user profiling methods do not discriminate the emotional differences of users of different genders and ages on social media against a background of multi-modality and a lack of domain sentiment labels. This paper adopts the sentiment analysis of images and text to improve label classification, incorporating gender and age differences in the sentiment analysis of multi-modal social-media user profiles. In the absence of domain sentiment labels, instance transfer learning technology is used to express the learning method with the sentiment of text and images; the semantic association learning of multi-modal data of graphics and text is realized; and a multi-modal attention mechanism is introduced to establish the hidden image and text. Alignment relationships are used to address the semantic and modal gaps between modalities. A multi-modal user portrait label classification model (MPCM) is constructed. In an analysis of the sentiment data of User users on Facebook, Twitter, and News, the MPCM method is compared with the naive Bayes, Latent Dirichlet Allocation, Tweet-LDA and LUBD-CM(3) methods in terms of accuracy, precision, recall and the FL-score. At a 95% confidence, the performance is improved by 1% to 4% by using the MPCM method.

The impact of economic uncertainty on carbon emission: Evidence from China

Economic uncertainty (EU) has received greater attention in recent years Based on media data, this research applies the Latent Dirichlet Allocation method to create a new provincial EU index and from panel data of provinces in China examines how EU impacts carbon emissions. The calculations illustrate that province-level EU in the country exerts a significantly positive influence on province-level carbon emissions. A mechanism test indicates that EU raises carbon emissions by harming green technology innovation, enterprise profits, and carbon emission efficiency. Companies hesitantly make decisions, put off innovations, and use cheaper, environmentally-unfriendly materials, and so a moderating test shows that a province's economic level suppresses the impact of EU on carbon emissions whereas foreign direct investment enhances it. Developed areas are more capable of bearing the negative effect from high EU, while foreign companies tend to focus on short-term profit instead of environment protection. Central and northeast regions of China suffer a significant impact from EU on carbon emissions, but the east and west regions incur no obvious change when undergoing EU. The west region is too backwards to see any impact from EU, while the east region is the opposite. Carbon emissions from transportation and industrial production incur a significant impact from EU, but not from daily life activities. Spatial analysis describes that carbon emissions see a significant impact of EU from other provinces. Governments and decision makers should thus give carbon emissions a higher priority, and the case of China is rather applicable to other regions and countries.

How recommendation letters affect career Mobility：Evidence from a social networking sites LinkedIn

Social networking sites (SNSs) provide users with ample opportunities to share their own information and participate in social browsing to get to know others. Drawing upon signaling theory, this paper investigated how and to what extent recommendation letters’ content information (breadth and depth) and external information (field and relationship) affect career mobility. Based on the open data from LinkedIn, this paper adopted the Latent Dirichlet Allocation (LDA) method to capture the breadth and depth of the recommendation letter, and used elaboration likelihood model to test our model. The results show that both content information and external information of the recommendation letter have significant positive impacts on the career mobility. In addition, the social network size shown converse moderating effects between recommendation letter content information and career mobility. This paper extends the boundaries of signaling theory by creatively using machine learning methods to measure textual signals (i.e., recommendation letter). The research results provide practical implications for personal career development and firm talent management.

Identifying Latent Topics and Trends in Premature Infant-Related Nursing Studies Using a Latent Dirichlet Allocation Method.

This study aimed to identify topics and within-topic core keywords in premature infant-related nursing studies published in Korean and international academic journals using topic modeling and to compare and analyze the trends in Korean and international studies. Journal databases were searched to extract nursing studies involving premature infants from 1998 to 2020. Journal databases included MEDLINE, Web of Science, CINAHL, and EMBASE for international studies and DBpia, the National Digital Science Library, the Korea Citation Index, and the Research Information Sharing Service for Korean studies. Abstracts from the selected 182 Korean and 2502 international studies were analyzed using NetMiner4.4.3e. In results, four similar topics (Korean vs international) were "pain intervention" versus "pain management"; "breast feeding practice" versus "breast feeding"; "kangaroo mother care"; and "parental stress" versus "stress & depression." Two topics that appeared only in the international studies were "infection management" and "oral feeding & respiratory care." Overall, the international studies dealt with diverse topics directly associated with premature. Korean studies mainly dealt with topics related to mothers of premature infants, whereas studies related to premature infants were insufficient. Nursing research in Korea needs to be expanded to research topics addressing premature infants.

Clustering and Topic Modeling of Verdicts of Narcotics Cases Using Machine Learning

Narcotics are a grave crime that can lead to addiction, loss of consciousness, and even death. Furthermore, narcotics can damage society’s environment. Narcotics criminal cases have been reported widely in Indonesia. The variety of narcotics cases makes it extremely difficult for judges to make decisions. Therefore, it is necessary to study and analyze the judge’s decisions from the data on the narcotics cases. In this study, we propose using a machine learning approach based on K-means clustering method for clustering and analyzing the verdicts on narcotics cases to see the trend of the verdicts on narcotics cases. In addition, we also use latent Dirichlet allocation (LDA) topic modeling to study the trend of these narcotics cases. Based on the results of the study using K=3 for clustering, there were three categories of verdicts: decisions with light sentences (less than three years), decisions with moderate sentences (three to six years), and decisions with severe sentences (more than ten years). Furthermore, using topic modeling based on the LDA method, the top three topics of narcotics cases based on the verdicts were determined, namely: the first topic refers to verdicts where narcotics perpetrators are found guilty; the second topic refers to verdicts with evidence of marijuana-type narcotics; and the third topic refers to verdicts with evidence of methamphetamine-type narcotics.

Revolution trend investigation of tourism destination image with machine learning

In tourism destination image research, topic modeling, mutual relationships among keywords, and image evolution trends have received limited attention. Filling these gaps in the extant literature, based on tourist gaze theory, this study is the first to use Latent Dirichlet Allocation, text visualization, and topic evolution trend methods to explore the Dunhuang Mogao Grottoes’ image. Analyzing 22,371 reviews posted on tourism social networks, we find that (1) prominence, uniqueness, a rich culture, convenience, and enormous influence are five dimensions of the Dunhuang Mogao Grottoes’ image; (2) these dimension keywords display different cooccurrence relationships; and (3) the dimension evolution trends vary in direction and degree. Our study enriches existing literature and provides important practical implications for the destination marketing field.

Topic Modelling of Web Pages with Latent Dirichlet Allocation Methods

Topic modelling with Latent Dirichlet Allocation (LDA) is a popular technique used in natural language processing to uncover hidden thematic structures within a collection of documents. When applied to web pages, LDA can help in identifying prevalent topics or themes across these pages.. This study delves into the utilization of Latent Dirichlet Allocation (LDA) methods to extract underlying topics within web pages, a fundamental pursuit in understanding the multifaceted landscape of online information. Web content analysis presents unique challenges owing to its diverse nature—comprising text, images, videos, and structured HTML elements—mandating rigorous preprocessing strategies to homogenize the data. By adapting the LDA model to accommodate these challenges, this research tackles the task of uncovering latent thematic structures prevalent across web content. Methodologically, the study explores parameter tuning and model adaptation to optimize LDA for web page analysis, navigating complexities such as varied content formats, noise, and inherent biases in web data. Addressing these intricacies involves parsing HTML, extracting meaningful textual information, and refining tokenization processes. Evaluating the fidelity and interpretability of discovered topics becomes pivotal, prompting the utilization of coherence scores, perplexity metrics, and human assessment to gauge the quality of generated topics. Additionally, this research confronts the dynamic nature of web content, proposing strategies like continuous model retraining and dynamic topic modeling to accommodate evolving trends and updates. Practical applications of the extracted topics span a spectrum of domains, encompassing content recommendation systems, user behavior analysis, sentiment analysis, targeted advertising, and the enhancement of search algorithms for improved relevance and user engagement. Supported by illustrative case studies, this study elucidates how LDA serves as a potent mechanism to distill coherent and meaningful topics from web pages, offering invaluable insights into the hidden structures within the vast expanse of online information. St This comprehensive abstract encapsulates the depth and breadth of employing LDA for the analysis of web content, encompassing challenges, methodologies, evaluations, applications, and real-world implications.

Exploring Blockchain Research in Supply Chain Management: A Latent Dirichlet Allocation-Driven Systematic Review

Blockchain technology has emerged as a tool with the potential to enhance transparency, trust, security, and decentralization in supply chain management (SCM). This study presents a comprehensive review of the interplay between blockchain technology and SCM. By analyzing an extensive dataset of 943 articles, our exploration utilizes the Latent Dirichlet Allocation (LDA) method to delve deep into the thematic structure of the discourse. This investigation revealed ten central topics ranging from blockchain’s transformative role in supply chain finance and e-commerce operations to its application in specialized areas, such as the halal food supply chain and humanitarian contexts. Particularly pronounced were discussions on the challenges and transformations of blockchain integration in supply chains and its impact on pricing strategies and decision-making. Visualization tools, including PyLDAvis, further illuminated the interconnectedness of these themes, highlighting the intertwined nature of blockchain adoption challenges with aspects such as traceability and pricing. Despite the breadth of topics covered, the paper acknowledges its limitations due to the fast-evolving nature of blockchain developments during and after our analysis period. Ultimately, this review provides a holistic academic snapshot, emphasizing both well-developed and nascent research areas and guiding future research in the evolving domain of blockchain in SCM.