Latent Topics Research Articles

Visualizing Change and Correlation of Topics With LDA and Agglomerative Clustering on COVID-19 Vaccine Tweets

More than three years have passed since the COVID-19 outbreak. The end of the pandemic is in sight, and even some countries have declared COVID-19 as endemic. During the pandemic, COVID-19 vaccination is essential to controlling the situation in many countries. Monitoring opinions on social media is important to understand the public reception of the COVID-19 vaccine. In this study, we combine LDA with agglomerative clustering to investigate the reactions of people from various countries during different periods for approximately one year after a few countries officially approved vaccination. Agglomerative clustering is performed on the latent topics based on the Hellinger distance for measuring the distance between the distributions of the latent topic. As a result, the proposed methods can show relative closeness among the topics and visualize the topic changes over time. The analysis shows that the significant focuses of tweets in the four phases with an interval of three months are the effectiveness of the vaccine, vaccine booster, Delta variant, and vaccine for children, which coincide with the timeline of the critical COVID events announced by WHO.

Question Tags or Text for Topic Modeling: Which is better

Topic modelling is a probabilistic based statistical model used to find the latent topics that best depicts the content of the documents. Community Question Answering websites such as Quora, Stack Overflow and Yahoo! Answers have been prevalently in use, performs topic modeling as lot of queries pour in on daily basis which make it challenging to understand, summarize and synthesize the main topic of discussions. On these websites there are basically two sources of information that are available to analyze the key latent topics: questions text and tags. Questions are in textual format and tags are the keywords or tokens that are related to the question being asked which describes the content of the question. In past studies, most of the researchers have used question text for the purpose of topic modeling. It is still unclear why tag is not being considered for topic modeling. To combat this issue, this paper performs topic modeling using both question tags and text. The topic modeling based on tags has been compared with text based on two metrics namely coherence and perplexity. Experiment has been conducted on three real time datasets namely Artificial intelligence, Software Engineering and quantum computing from Stack exchange website. At high level tag-based topic modelling looked promising but closer observation revealed the opposite. It has been found that topic modeling using question text is preferable as topic modelling using tags collapses after a certain number of topics.

A conversação pública acerca da vacina e da vacinação contra covid-19 no Twitter: um estudo infodemiológico

Abstract The present research is infodemiological and aims to characterize the public conversation about the vaccine and vaccination against COVID-19 in Brazil that took place on Twitter. To map the latent topics in the collected data (13.2 million tweets), topic modeling was used. since the effectiveness of public vaccination policies is conditioned by its population adherence. From the results, 50 topics were identified and discussed over the period of analysis, 11 of which have direct potential to promote vaccine hesitancy, categorized into thematic groups, namely Accountability, Public Policies, Conflicts, Bioethics, Experience Reports, Economic Aspects, Conspiracy Theories; and Safety, Efficacy and Importance of Vaccines. There was a strong politicization translated by the most retweeted profiles and hashtags used, not only partisan, but mostly civic, represented by users’ participation through the manifestation of social control in health.

Advanced Hierarchical Topic Labeling for Short Text

Hierarchical Topic Modeling is the probabilistic approach for discovering latent topics distributed hierarchically among the documents. The distributed topics are represented with the respective topic terms. An unambiguous conclusion from the topic term distribution is a challenge for readers. The hierarchical topic labeling eases the challenge by facilitating an individual, appropriate label for each topic at every level. In this work, we propose a BERT-embedding inspired methodology for labeling hierarchical topics in short text corpora. The short texts have gained significant popularity on multiple platforms in diverse domains. The limited information available in the short text makes it difficult to deal with. In our work, we have used three diverse short text datasets that include both structured and unstructured instances. Such diversity ensures the broad application scope of this work. Considering the relevancy factor of the labels, the proposed methodology has been compared against both automatic and human annotators. Our proposed methodology outperformed the benchmark with an average score of 0.4185, 49.50, and 49.16 for cosine similarity, exact match, and partial match, respectively.

Detecting latent topics and trends of digital twins in healthcare: A structural topic model-based systematic review.

Digital twins (DTs) have received widespread attention recently, providing new ideas and possibilities for future healthcare. This review aims to provide a quantitative review to analyze specific study contents, research focus, and trends of DT in healthcare. Simultaneously, this review intends to expand the connotation of "healthcare" into two directions, namely "Disease treatment" and "Health enhancement" to analyze the content within the "DT + healthcare" field thoroughly. A data mining method named Structure Topic Modeling (STM) was used as the analytical tool due to its topic analysis ability and versatility. Google Scholar, Web of Science, and China National Knowledge Infrastructure supplied the material papers in this review. A total of 94 high-quality papers published between 2018 and 2022 were gathered and categorized into eight topics, collectively covering the transformative impact across a broader spectrum in healthcare. Three main findings have emerged: (1) papers published in healthcare predominantly concentrate on technology development (artificial intelligence, Internet of Things, etc.) and application scenarios(personalized, precise, and real-time health service); (2) the popularity of research topics is influenced by various factors, including policies, COVID-19, and emerging technologies; and (3) the preference for topics is diverse, with a general inclination toward the attribute of "Health enhancement." This review underscores the significance of real-time capability and accuracy in shaping the future of DT, where algorithms and data transmission methods assume central importance in achieving these goals. Moreover, technological advancements, such as omics and Metaverse, have opened up new possibilities for DT in healthcare. These findings contribute to the existing literature by offering quantitative insights and valuable guidance to keep researchers ahead of the curve.

Sentiment Analysis using Latent Dirichlet Allocation for Aspect Term Extraction

This work proposes a sentiment analysis approach for decision-making in product design, analysis, and market share. The approach incorporates user-generated text data in the form of consumer reviews to extract product features using topic-based modeling methods. Latent Dirichlet Allocation (LDA) is employed to extract aspect categories from the data and identify the sentiment of each review using the VADER sentiment analyzer. The performance of the proposed method is evaluated in terms of accuracy, with an achieved result of 80%. The extracted topics are also summarized to provide leads for product design and quality assurance. The approach can be used by manufacturers, retailers, and suppliers to understand customers' opinions about their products better and make better decisions. LDA is a powerful unsupervised method that can extract latent topics from a collection of documents; this method has been widely used in text mining, information retrieval, and natural language processing. The accuracy can be improved by using more sophisticated models or more data.

Application of Artificial Intelligence to improve Customer Understanding: Transformer based Topic Modeling in practice

Since the past few years, so-called pre-trained Language Models (PLM) are considered state-of-the-art in the field of Natural Language Processing (NLP) and are thus experiencing widespread and successful application. In addition to traditional supervised Machine Learning (ML) tasks such as spam email or customer churn classification, this technology opens up advanced approaches to unsupervised learning and data analytics in general. One of particular interest is the automatic identification of latent topics within a large collection of texts, also known as Topic Modeling (TM). Such modelling approaches offer great potential, especially for industrial environments as well as the consumer goods market, to explore increasing amounts of data from diverse and constantly growing data sources. As a holistic concept, it can be utilized in a highly targeted and efficient manner for applications through an optimized combination of Artificial Intelligence (AI) and Cloud Computing (CC) systems.

GCN recommendation model based on the fusion of dynamic multiple-view latent interest topics

GCN recommendation model based on the fusion of dynamic multiple-view latent interest topics

Comparison of Methods for Estimating Temporal Topic Models From Primary Care Clinical Text Data: Retrospective Closed Cohort Study

Health care organizations are collecting increasing volumes of clinical text data. Topic models are a class of unsupervised machine learning algorithms for discovering latent thematic patterns in these large unstructured document collections. We aimed to comparatively evaluate several methods for estimating temporal topic models using clinical notes obtained from primary care electronic medical records from Ontario, Canada. We used a retrospective closed cohort design. The study spanned from January 01, 2011, through December 31, 2015, discretized into 20 quarterly periods. Patients were included in the study if they generated at least 1 primary care clinical note in each of the 20 quarterly periods. These patients represented a unique cohort of individuals engaging in high-frequency use of the primary care system. The following temporal topic modeling algorithms were fitted to the clinical note corpus: nonnegative matrix factorization, latent Dirichlet allocation, the structural topic model, and the BERTopic model. Temporal topic models consistently identified latent topical patterns in the clinical note corpus. The learned topical bases identified meaningful activities conducted by the primary health care system. Latent topics displaying near-constant temporal dynamics were consistently estimated across models (eg, pain, hypertension, diabetes, sleep, mood, anxiety, and depression). Several topics displayed predictable seasonal patterns over the study period (eg, respiratory disease and influenza immunization programs). Nonnegative matrix factorization, latent Dirichlet allocation, structural topic model, and BERTopic are based on different underlying statistical frameworks (eg, linear algebra and optimization, Bayesian graphical models, and neural embeddings), require tuning unique hyperparameters (optimizers, priors, etc), and have distinct computational requirements (data structures, computational hardware, etc). Despite the heterogeneity in statistical methodology, the learned latent topical summarizations and their temporal evolution over the study period were consistently estimated. Temporal topic models represent an interesting class of models for characterizing and monitoring the primary health care system.

Knowledge Trajectories Detection and Prediction of Modern Emergency Management in China Based on Topic Mining from Massive Literature Text

China has witnessed dramatic advances in emergency management in the past two decades, while the knowledge trajectories and future trends of related research are still unclear. This study takes the published articles in China National Knowledge Infrastructure as a data sample and introduces text mining and machine learning methods, namely Latent Dirichlet Allocation combined with the Hidden Markov Model, to detect and predict the knowledge trajectories of Chinese modern emergency management research. We analyzed 5180 articles, equivalent to approximately 1,110,000 Chinese characters, from 2003 to 2021, and mined 35 latent research topics. By labeling the topics manually and analyzing the evolutionary hotspots, confusion and transition features, and transition direction and network of the topics, we explored the knowledge trajectories of emergency management research in China. By training the HMM model, we predicted the research trends in the next five years. The main conclusions are: a mapping relationship exists between the hotspots of the published articles and the main events of emergency management in China; most emergency management research topics could confuse and transfer with others in the evolution process, and seven significant paths exist in the transition network. The research topics in the following years will be more detailed and concerned with the intellectual needs of modernization.

Exploring mobile banking adoption and service quality features through user-generated content: the application of a topic modeling approach to Google Play Store reviews

PurposeThe primary purpose of this research is to analyze the online user reviews, where real customer experiences can be observed, with text mining and machine learning approaches, which are seen as a gap in the related literature. This study aims to compare the latent themes uncovered by the topic modeling approach with studies focused on both mobile banking (m-banking) adaptation and service quality features, suggest new aspects and examine the effect of latent topics on customer satisfaction.Design/methodology/approachThis study analyzed 21,526 reviews posted by customers of private and state banks operating in Türkiye. An unsupervised machine learning method, Latent Dirichlet algorithm (LDA), was conducted to reveal topics, and the distribution of all reviews was visualized with the t-SNE algorithm. Random Forest, logistic regression, k-nearest neighbors (kNN) and Naive Bayes algorithms were utilized to predict user satisfaction through the given score.FindingsIn total, 11 topics were revealed by considering user reviews based on their experience. Among these topics, perceived usefulness and convenience and time-saving are much more important in the scoring given to m-banking apps. Furthermore, in more detail, seven topics have been identified related to technical and security problems related to m-banking apps.Originality/valueThis paper is a pioneer study regarding the method used and sample size reached in the m-banking literature. The findings also provide fresh insight into the post-Covid-19 era, both academically and practically, by providing new features for mobile bank adoption.

Predication of Writing Originality Based on Computational Linguistics.

Existing assessment methods of writing originality have been criticized for depending heavily on subjective scoring methods. This study attempted to investigate the use of topic analysis and semantic networks in assessing writing originality. Written material was collected from a Chinese language test administered to eighth-grade students. Two steps were performed: 1. Latent topics of essays in each writing task were identified, and essays on the same topic were treated as a refined reference group, within which an essay was to be evaluated; 2. A group of features was developed, including four categories, i.e., path distance, semantic differences, centrality, and similarity of the network drawn from each text response, which were used to quantify the differences among essays. The results show that writing originality scoring is not only related to the intrinsic characteristics of the text, but is also affected by the reference group in which it is to be evaluated. This study proves that computational linguistic features can be a predictor of originality in Chinese writing. Each feature type of the four categories can predict originality, although the effect varies across various topics. Furthermore, the feature analysis provided evidence and insights to human raters for originality scoring.

Automatic topics extraction from crowdsourced cyclists near-miss and collision reports using text mining and Artificial Neural Networks

Cycling is an eco-friendly and sustainable mode of transportation. Despite its benefits, the cyclists’ risk of collision is still high when interacting with other road users. This study analyzed self-reported near-miss and collision event descriptions for the United States provided by the crowdsourcing platform, BikeMaps.org. Innovative and efficient analytic methods are needed to generate useful information from unstructured textual data sources in the transportation domain. In this study, explorative text mining, topic modeling, and machine learning are utilized to gain insights from the unstructured textual descriptions of crowdsourced near-miss and collision events. The approaches are used to unveil prevalent words and word associations for near-miss and collision events. Structural Topic Modeling (STM) is deployed to autogenerate latent themes or topics from the event descriptions. The generated topic proportions are used as input in Artificial Neural Networks (ANN) to estimate the cyclist’s propensity to a collision. It was found that cyclists had a higher propensity to a collision in topics that articulated vehicle encroachment to the bike lane, on-street parking close or into the bike lane resulting in dooring, and drivers’ violations at the crosswalk. The results and methodology used in this study can assist engineers, policymakers, and law enforcement officers to proactively reduce potential cyclist collisions, prioritizing areas where cyclist safety improvements are needed, and ultimately promoting bicycle ridership in our communities.

Latent Dirichlet allocation topic modeling of free-text responses exploring the negative impact of the early COVID-19 pandemic on research in nursing.

To derive latent topics from free-text responses on the negative impact of the pandemic on research activities and determine similarities and differences in the resulting themes between academic-based and clinical-based researchers. We performed a secondary analysis of free-text responses from a cross-sectional online survey conducted by the Japan Academy of Nursing Science of its members in early 2020. The participants were categorized into two groups by workplace (academic-based and clinical-based researchers). Latent Dirichlet allocation (LDA) topic modeling was used to extract latent topics statistically and list important keywords/text associated with the topics. After organizing similar topics by principal component analysis (PCA), we finally derived topic-associated themes by reading the keywords/texts and determining the similarity and differences of the themes between the two groups. A total of 201 respondents (163 academic-based and 38 clinical-based researchers) provided free-text responses. LDA identified eight and three latent topics for the academic-based and clinical-based researchers, respectively. While PCA re-grouped the eight topics derived from the former group into four themes, no merging of the topics from the latter group was performed resulting in three themes. The only theme common to the two groups was "barriers to conducting research," with the remaining themes differing between the groups. Using LDA topic modeling with PCA, we identified similarities and differences in the themes described in free-text responses about the negative impact of the pandemic between academic-based and clinical-based researchers. Measures to mitigate the negative impact of pandemics on nursing research may need to be tailored separately.

A bibliometric review on latent topics and research trends in the growth mindset literature for mathematics education.

Embracing a growth mindset is essential to students' academic improvement. This manuscript aims to better understand the existing literature on the role and effects of the growth mindset in mathematics teaching and learning. It provides an updated perspective on the research regarding the growth mindset in mathematics education. The dataset comprises 85 journal articles published from 2012 to 2022 retrieved from the Web of Science (WOS) and Scopus databases. The current study applies a methodology based on bibliometric analysis techniques. The analysis reveals and corroborates several patterns from the research trends, journals, countries, and authors that have significant impacts on the research field. The findings show that USA, UK, and Norway are the most productive countries in publishing research on the topic. Moreover, the results of the thematic analysis indicate that the topics discussed among most of the articles in the dataset include engagement, implementation, persistence, children, fluid intelligence, and skills. The longitudinal trends in research themes based on study keywords illustrate an evolution in the research from the concept of mindsets to implicit theories on the growth mindset alongside academic achievement. Lastly, this study also provides an overview of the conceptual structure underlying studies on the growth mindset, which offers valuable insights into potential research topics for academics and practitioners seeking to explore the growth mindset in the future.

Extracting information and inferences from a large text corpus.

The usage of various software applications has grown tremendously due to the onset of Industry 4.0, giving rise to the accumulation of all forms of data. The scientific, biological, and social media text collections demand efficient machine learning methods for data interpretability, which organizations need in decision-making of all sorts. The topic models can be applied in text mining of biomedical articles, scientific articles, Twitter data, and blog posts. This paper analyzes and provides a comparison of the performance of Latent Dirichlet Allocation (LDA), Dynamic Topic Model (DTM), and Embedded Topic Model (ETM) techniques. An incremental topic model with word embedding (ITMWE) is proposed that processes large text data in an incremental environment and extracts latent topics that best describe the document collections. Experiments in both offline and online settings on large real-world document collections such as CORD-19, NIPS papers, and Tweet datasets show that, while LDA and DTM is a good model for discovering word-level topics, ITMWE discovers better document-level topic groups more efficiently in a dynamic environment, which is crucial in text mining applications.

Tiered sentence based topic model for multi-document summarization

In this work, a probabilistic two level topic model named Tiered Sentence based Topic Model is proposed which models the document at sentence and word levels and infer hierarchical latent topics for sentences. The proposed model uses two latent variables for the generation of words- a super topic and a subtopic for each sentence of the document, to model word groupings at sentence level. Popular super topics identify general theme of the documents and are used for selecting summary sentences. The model parameters are used for ranking sentences considering sentence importance and topic coverage. Collapsed Gibbs sampling is used for inference and parameter estimation. The proposed model is used to compare with two sentence based topic models- SenLDA and LDCC on query focused multi-document summarization task, over standard DUC2005 dataset using ROUGE-1 and ROUGE-2 precision and recall scores. The proposed model performs better than Latent Dirichlet Allocation and SenLDA but has been outperformed by LDCC.

LDA-based topic modeling for COVID-19-related sports research trends.

The COVID-19 pandemic could generate a turning point for introducing a new system for sports participation and business. The purpose of this study is to explore trends and topic structures of COVID-19-related sports research by analyzing the relevant literature. Sports studies related to COVID-19 were collected in searching international academic databases. After the pre-processing step using the refinement and morpheme analysis function of the Net Miner program, topic modeling and social network analysis were used to analyze Journal Citation Reports found using the search term 'COVID-19 sports'. As a result, this study used subject modeling to reveal important potential topics in COVID-19-related sports research articles. 'Sports participation', 'elite players', and 'sports industry' were macroscopically classified, and detailed research topics could be identified from each division. This study revealed important latent topics from COVID-19-related sports research articles using topic modeling. The results of the research elucidate the structure of academic knowledge on this topic and provide guidance for future research.

A supervised topic embedding model and its application.

We propose rTopicVec, a supervised topic embedding model that predicts response variables associated with documents by analyzing the text data. Topic modeling leverages document-level word co-occurrence patterns to learn latent topics of each document. While word embedding is a promising text analysis technique in which words are mapped into a low-dimensional continuous semantic space by exploiting the local word co-occurrence patterns within a small context window. Recently developed topic embedding benefits from combining those two approaches by modeling latent topics in a word embedding space. Our proposed rTopicVec and its regularized variant incorporate regression into the topic embedding model to model each document and a numerical label paired with the document jointly. In addition, our models yield topics predictive of the response variables as well as predict response variables for unlabeled documents. We evaluated the effectiveness of our models through experiments on two regression tasks: predicting stock return rates using news articles provided by Thomson Reuters and predicting movie ratings using movie reviews. Results showed that the prediction performance of our models was more accurate in comparison to three baselines with a statistically significant difference.

Topic modeling and sentiment analysis of Chinese people’s attitudes toward volunteerism amid the COVID-19 pandemic

The COVID-19 pandemic has created an urgent need for volunteers to complement overwhelmed public health systems. This study aims to explore Chinese people's attitudes toward volunteerism amid the COVID-19 pandemic. To this end, we identify the latent topics in volunteerism-related microblogs on Weibo, the Chinese equivalent of Twitter using the topic modeling analysis via Latent Dirichlet Allocation (LDA). To further investigate the public sentiment toward the topics generated by LDA, we also conducted sentiment analysis on the sample posts using the open-source natural language processing (NLP) technique from Baidu. Through an in-depth analysis of 91,933 Weibo posts, this study captures 10 topics that are, in turn, distributed into five factors associated with volunteerism in China as motive fulfillment (n = 31,661, 34.44%), fear of COVID-19 (n = 22,597, 24.58%), individual characteristic (n = 17,688, 19.24%), government support (n = 15,482, 16.84%), and community effect (n = 4,505, 4.90%). The results show that motive fulfillment, government support, and community effect are the factors that could enhance positive attitudes toward volunteerism since the topics related to these factors report high proportions of positive emotion. Fear of COVID-19 and individual characteristic are the factors inducing negative sentiment toward volunteerism as the topics related to these factors show relatively high proportions of negative emotion. The provision of tailored strategies based on the factors could potentially enhance Chinese people's willingness to participate in volunteer activities during the COVID-19 pandemic.