Search

The authors develop an attribute-based mixed-membership model of consumers' preference for stockkeeping units in store assortments. The model represents the underlying "topics of interest" that drive shopping behaviors as probability distributions over product attributes. It overcomes several limitations of latent Dirichlet allocation topic models and is particularly useful for making preference predictions in large and frequently changing store assortments. The authors apply the proposed model to investigate topics driving browsing and purchase activities in an online deal marketplace of fashion products and explore how preference structures evolve over time. They find commonalities and differences in the topics that drive the browsing and purchase stages of online shopping processes. In general, browsing covers a broader range of product attributes than purchases. Consumers tend to browse products of premium positioning and/or deep discounts in the deal marketplace, but when purchasing, they tend to gravitate toward lower-tiered products at their original prices and modest depths of discounts. The authors illustrate how insights from the proposed model can be utilized to profile consumers based on their price preferences and to improve personalized product recommendations. They show that the model's performance is particularly strong in predicting preferences for new products that are not in the existing assortment. [ABSTRACT FROM AUTHOR]

Social networks are the most important stage in achieving big data in the discipline of health, with tweets from anywhere in the world currently enabling. It's also a great source of information for researching health issues and forecasting healthcare decisions. In terms of revenue and service, health care has become one of the world's leading industries. Every day, billions of customers use Twitter to share health-related topics, their views, and opinions on various healthcare topics. By aggregating millions of users' self-reported health conditions, we hope to distinguish the variety of health information discussed on Twitter. A topic modelling framework for discovering health topics on Twitter, a social networking website, is described. This is an exploratory approach to learning about the health topics that are frequently discussed on social media. In this sense, the required visualisations are a necessary measure to trim the information in order to identify the cluster's direction. The proposed topic model, i.e., Distributed Semantic Assisted Non-Negative matrix factorization, contributes to the pool of health statistics by balancing and pruning towards the direction of health topics from various perspective data sources. This assistance briefly describes the country's good public health structure and tracks the evolution of the main health-related tweets for early public advice. [ABSTRACT FROM AUTHOR]

We propose a novel visual analytics approach based on the Latent Dirichlet Allocation (LDA) model for exploring and interpreting the dynamic evolution of networks. In this approach, we define networks as documents and relationships within networks as words. Using this definition, the LDA model is able to extract a list of structures that fuse relationships and connect the network features. We project networks described by the extracted structures with probabilistic assignments as points into a two-dimensional space via dimensionality reduction techniques. Users can identify evolution states in dynamic networks, including stable states, recurrent states, outlier states, and state transitions. To facilitate the interpretation of evolution states, we provide a novel small multiples view that shows how the extracted structures behave over time. We demonstrate the effectiveness of our work through case studies conducted on two real-world dynamic networks. [ABSTRACT FROM AUTHOR]

Topic models are mathematical and statistical models used to analyze textual data. The objective of topic models is to gain information about the latent semantic space of a set of related textual data. The semantic space of a set of textual data contains the relationship between documents and words and how they are used. Topic models are becoming more common in educational measurement research as a method for analyzing students' responses to constructed-response items. Two popular topic models are latent semantic analysis (LSA) and latent Dirichlet allocation (LDA). LSA uses linear algebra techniques, whereas LDA uses an assumed statistical model and generative process. In educational measurement, LSA is often used in algorithmic scoring of essays due to its high reliability and agreement with human raters. LDA is often used as a supplemental analysis to gain additional information about students, such as their thinking and reasoning. This article reviews and compares the LSA and LDA topic models. This article also introduces a methodology for comparing the semantic spaces obtained by the two models and uses a simulation study to investigate their similarities. [ABSTRACT FROM AUTHOR]

Student evaluations of teaching (SET) have been examined over the years to better understand the student experience, with an increasing portion of literature exploring the presence of implicit bias in SET surveys against minority gender and ethnic lecturers. This study explores free-text comments made by students from a large public university in Australia over the period 2010-2016, using a semi-supervised statistical approach. Data were collected via surveys administered online at the end of each course to every student officially enrolled in that course via the learning management system, and completion of the surveys was voluntary. We build a probabilistic topic model which incorporates student and lecturer characteristics into the topic formation process. We make statistical inference on the effects of gender and cultural or language backgrounds based on the topic and word prevalence probabilities. The results showed clear separation of topics discussed betweenmale and female lecturers. From a gendered perspective, our topic analyses have found that students are significantly more likely to critique female lecturers to improve on structural aspects of the course, as well as aspects of time management and control of the lecturing environment. In comparison, male lecturers were significantlymore likely to be critiqued on specific aspects related to lecture delivery. Lecturers from non-English speaking backgrounds were more likely to be both critiqued and praised for the clarity of their delivery. [ABSTRACT FROM AUTHOR]

Abstract The adoption of high-performance computing (HPC) in healthcare has gained significant attention in recent years, driving advancements in medical research and clinical practice. Exploring the literature on HPC implementation in healthcare is valuable for decision-makers as it provides insights into potential areas for further investigation and investment. However, manually analyzing the vast number of scholarly articles is a challenging and time-consuming task. Fortunately, topic modeling techniques offer the capacity to process extensive volumes of scientific literature, identifying key trends within the field. This paper presents an automatic literature analysis framework based on a state-of-art vector-based topic modeling algorithm with multiple embedding techniques, unveiling the research trends surrounding HPC utilization in healthcare. The proposed pipeline consists of four phases: paper extraction, data preprocessing, topic modeling and outlier detection, followed by visualization. It enables the automatic extraction of meaningful topics, exploration of their interrelationships, and identification of emerging research directions in an intuitive manner. The findings highlight the transition of HPC adoption in healthcare from traditional numerical simulation and surgical visualization to emerging topics such as drug discovery, AI-driven medical image analysis, and genomic analysis, as well as correlations and interdisciplinary connections among application domains.

Graph Convolutional Network (GCN) has been extensively studied in the task of short text classification (STC), utilizing global graphs that incorporate texts at different levels of granularity to learn text embeddings. However, the GCN-based methods only focus on the alignment between ground-truth labels and predicted labels, overlooking the geometric structure implicitly encoded by the graph. To address this limitation, we propose a novel GCN-based method that is entitled Topic-aware Cosine GCN (ToCo-GCN) for the STC. The ToCo-GCN defines and captures underlying geometric structures of short texts from different categories in the cosine space. Specifically, the ToCo-GCN regards the within-class and between-class geometric structures as constraint, aiming to learn both representative and discriminative short text representations. Moreover, to mitigate the inherent sparsity problem of short texts, the ToCo-GCN augment the text graph with latent topics. Experimental results on 8 STC datasets demonstrate that the ToCo-GCN is superior to state-of-the-art baselines in terms of Accuracy and Macro-F1 score. [ABSTRACT FROM AUTHOR]

Although public diplomacy and its influence on foreign public opinion have been central themes in recent research, the latter often lacks methodological diversity and does not consider how states create competing images. Our study offers a framework for understanding Italian public opinion using a triangulation approach that combines traditional public opinion surveys with advanced text analysis of social media content. We compare the representation of China and the United States in the Italian Twitter community during the first wave of the COVID-19 pandemic. Our analysis shows that Italians who perceived China as a viable alternative to Western governments in public opinion surveys did so because they distrusted Western leaders and institutions. [ABSTRACT FROM AUTHOR]

The adoption of high-performance computing (HPC) in healthcare has gained significant attention in recent years, driving advancements in medical research and clinical practice. Exploring the literature on HPC implementation in healthcare is valuable for decision-makers as it provides insights into potential areas for further investigation and investment. However, manually analyzing the vast number of scholarly articles is a challenging and time-consuming task. Fortunately, topic modeling techniques offer the capacity to process extensive volumes of scientific literature, identifying key trends within the field. This paper presents an automatic literature analysis framework based on a state-of-art vector-based topic modeling algorithm with multiple embedding techniques, unveiling the research trends surrounding HPC utilization in healthcare. The proposed pipeline consists of four phases: paper extraction, data preprocessing, topic modeling and outlier detection, followed by visualization. It enables the automatic extraction of meaningful topics, exploration of their interrelationships, and identification of emerging research directions in an intuitive manner. The findings highlight the transition of HPC adoption in healthcare from traditional numerical simulation and surgical visualization to emerging topics such as drug discovery, AI-driven medical image analysis, and genomic analysis, as well as correlations and interdisciplinary connections among application domains. [ABSTRACT FROM AUTHOR]

Student evaluations of teaching (SET) have been examined over the years to better understand the student experience, with an increasing portion of literature exploring the presence of implicit bias in SET surveys against minority gender and ethnic lecturers. This study explores free-text comments made by students from a large public university in Australia over the period 2010–2016, using a semi-supervised statistical approach. Data were collected via surveys administered online at the end of each course to every student officially enrolled in that course via the learning management system, and completion of the surveys was voluntary. We build a probabilistic topic model which incorporates student and lecturer characteristics into the topic formation process. We make statistical inference on the effects of gender and cultural or language backgrounds based on the topic and word prevalence probabilities. The results showed clear separation of topics discussed between male and female lecturers. From a gendered perspective, our topic analyses have found that students are significantly more likely to critique female lecturers to improve on structural aspects of the course, as well as aspects of time management and control of the lecturing environment. In comparison, male lecturers were significantly more likely to be critiqued on specific aspects related to lecture delivery. Lecturers from non-English speaking backgrounds were more likely to be both critiqued and praised for the clarity of their delivery.

We propose and evaluate an automated pipeline for discovering significant topics from legal decision texts by passing features synthesized with topic models through penalized regressions and post-selection significance tests. The method identifies case topics significantly correlated with outcomes, topic-word distributions which can be manually interpreted to gain insights about significant topics, and case-topic weights which can be used to identify representative cases for each topic. We demonstrate the method on a new dataset of domain name disputes and a canonical dataset of European Court of Human Rights violation cases. Topic models based on latent semantic analysis as well as language model embeddings are evaluated. We show that topics derived by the pipeline are consistent with legal doctrines in both areas and can be useful in other related legal analysis tasks. This article is part of the theme issue 'A complexity science approach to law and governance'. [ABSTRACT FROM AUTHOR]

Latent Dirichlet allocation (LDA) is one of the major models used for topic modelling. A number of models have been proposed extending the basic LDA model. There has also been interesting research to replace the Dirichlet prior of LDA with other pliable distributions like generalized Dirichlet, Beta‐Liouville and so forth. Owing to the proven efficiency of using generalized Dirichlet (GD) and Beta‐Liouville (BL) priors in topic models, we use these versions of topic models in our paper. Furthermore, to enhance the support of respective topics, we integrate mixture components which gives rise to generalized Dirichlet mixture allocation and Beta‐Liouville mixture allocation models respectively. In order to improve the modelling capabilities, we use variational inference method for estimating the parameters. Additionally, we also introduce an online variational approach to cater to specific applications involving streaming data. We evaluate our models based on its performance on applications related to text classification, image categorization and genome sequence classification using a supervised approach where the labels are used as an observed variable within the model. [ABSTRACT FROM AUTHOR]

The COVID-19 pandemic has had a profound impact on society and many governments and politicians were required to make quick decisions to protect public health. In reaching these decisions they needed to weigh advice from experts in many fields and then "sell" these decisions. However little was known at the time about the desires and aspirations of their electorates. Here we identify one such source of public opinion, signatures to various COVID-19 related Parliamentary e-petitions Whilst there were over 1,500 such e-petitions, we identify 27 common and meaningful topics and show how the strength of these topics varies between individual parliamentary constituencies. Using a measure of support for each topic in the constituencies we identify five constituency groupings. There are two Conservative voting groups, "Middle England" and "Equity of Support", the former are located largely in southern England whilst the latter are in the midlands and northern England. The groups that show strong support in Labour voting constituencies are ones around education and funding issues. This article shows how a targeted range of e-petitions can be grouped into topics and the popularity of topics established, and thereby function as a useful way of augmenting democracy and democratic institutions. [ABSTRACT FROM AUTHOR]

Optimization of medication therapy depends on maximizing benefits and minimizing side effects of medications. This research showed how a joint approach using text mining, natural language processing, and machine learning can provide information for personalized and optimized medication therapy. Reviews on the benefits and side effects of prescription and over-the-counter medications were used to determine how well an integrated supervised and unsupervised learning could learn medication satisfaction. Supervised learning with naïve-Bayes, non-linear support vector machine with radial basis function kernels, and random forests with CART decision trees was measured by a micro-aggregated Matthews correlation coefficient and a macro-averaged F1 measure. Random forests outperformed support vector machines by almost 250% and naive-Bayes by 600% on the two evaluation metrics. All models did better with three rating levels, instead of five. Topic modeling and stacked cluster analysis were coupled with parts-of-speech tagging and text mining operations to establish a robust data preprocessing procedure to eliminate noisy features from the data. Unsupervised topic modeling and clustering represented an exploratory validation of how easy supervised classification would be. Well-defined latent topics were discovered including topics on “sleep quality”, “the opportunity to get back to work”, and “weight gain”. Overlapping clusters revealed that incorporating more information on social, demographic, or medical history variables could improve classifier performance. This research provided evidence that medication satisfaction can be learned with carefully designed joint supervised, unsupervised, and natural language learning techniques.

Abstract Background In recent years, the discovery of clinical pathways (CPs) from electronic medical records (EMRs) data has received increasing attention because it can directly support clinical doctors with explicit treatment knowledge, which is one of the key challenges in the development of intelligent healthcare services. However, the existing work has focused on topic probabilistic models, which usually produce treatment patterns with similar treatment activities, and such discovered treatment patterns do not take into account the temporal process of patient treatment which does not meet the needs of practical medical applications. Methods Based on the assumption that CPs can be derived from the data of EMRs which usually record the treatment process of patients, this paper proposes a new CPs mining method from EMRs, an extended form of the traditional topic model - the temporal topic model (TTM). The method can capture the treatment topics and the corresponding treatment timestamps for each treatment day. Results Experimental research conducted on a real-world dataset of patients’ hospitalization processes, and the achieved results demonstrate the applicability and usefulness of the proposed methodology for CPs mining. Compared to existing benchmarks, our model shows significant improvement and robustness. Conclusion Our TTM provides a more competitive way to mine potential CPs considering the temporal features of the EMR data, providing a very prospective tool to support clinical diagnostic decisions.

The author proposes a topic model tailored to the study of creative documents (e.g., academic papers, movie scripts), which extends Poisson factorization in two ways. First, the creativity literature emphasizes the importance of novelty in creative industries. Accordingly, this article introduces a set of residual topics that represent the portion of each document that is not explained by a combination of common topics. Second, creative documents are typically accompanied by summaries (e.g., abstracts, synopses). Accordingly, the author jointly models the content of creative documents and their summaries, and captures systematic variations in topic intensities between the documents and their summaries. This article validates and illustrates the model in three domains: marketing academic papers, movie scripts, and TV show closed captions. It illustrates how the joint modeling of documents and summaries provides some insight into how people summarize creative documents and enhances understanding of the significance of each topic. It shows that the model described produces new measures of distinctiveness that can inform the perennial debate on the relation between novelty and success in creative industries. Finally, the author shows how the proposed model may form the basis for decision support tools that assist people in writing summaries of creative documents. [ABSTRACT FROM AUTHOR]

Background: In recent years, the discovery of clinical pathways (CPs) from electronic medical records (EMRs) data has received increasing attention because it can directly support clinical doctors with explicit treatment knowledge, which is one of the key challenges in the development of intelligent healthcare services. However, the existing work has focused on topic probabilistic models, which usually produce treatment patterns with similar treatment activities, and such discovered treatment patterns do not take into account the temporal process of patient treatment which does not meet the needs of practical medical applications. Methods: Based on the assumption that CPs can be derived from the data of EMRs which usually record the treatment process of patients, this paper proposes a new CPs mining method from EMRs, an extended form of the traditional topic model - the temporal topic model (TTM). The method can capture the treatment topics and the corresponding treatment timestamps for each treatment day. Results: Experimental research conducted on a real-world dataset of patients' hospitalization processes, and the achieved results demonstrate the applicability and usefulness of the proposed methodology for CPs mining. Compared to existing benchmarks, our model shows significant improvement and robustness. Conclusion: Our TTM provides a more competitive way to mine potential CPs considering the temporal features of the EMR data, providing a very prospective tool to support clinical diagnostic decisions. [ABSTRACT FROM AUTHOR]

Integrated writing assessments are tasks that require students to construct an essay using evidence from source texts. Topic models are statistical models used toanalyze text. This study analyzes two integrated writing tasks from students in grades 6 (N=1,250) and 8 (N=1,250) using topic models. Independent evaluations of the essays were also conducted to help generalize the results across the two grades and better understand students' underlying writing techniques. Four main categories were identified: (1) use of words that appeared in the prompt, (2) words that appeared in the source texts, (3) words that reflected inference, and (4) words that reflected integration. Higher-scoring students used words that reflect inference and integration of source text more often than lower-performing students. [ABSTRACT FROM AUTHOR]

Probabilistic Latent Semantic Analysis (PLSA) is a fundamental text analysis technique that models each word in a document as a sample from a mixture of topics. PLSA is the precursor of probabilistic topic models including Latent Dirichlet Allocation (LDA). PLSA, LDA and their numerous extensions have been successfully applied to many text mining and retrieval tasks. One important extension of LDA is supervised LDA (sLDA), which distinguishes itself from most topic models in that it is supervised. However, to the best of our knowledge, no prior work extends PLSA in a similar manner sLDA extends LDA by jointly modeling the contents and the responses of documents. In this paper, we propose supervised PLSA (sPLSA) which can efficiently infer latent topics and their factorized response values from the contents and the responses of documents. The major challenge lies in estimating a document's topic distribution which is a constrained probability that is dictated by both the content and the response of the document. To tackle this challenge, we introduce an auxiliary variable to transform the constrained optimization problem to an unconstrained optimization problem. This allows us to derive an efficient Expectation and Maximization (EM) algorithm for parameter estimation. Compared to sLDA, sPLSA converges much faster and requires less hyperparameter tuning, while performing similarly on topic modeling and better in response factorization. This makes sPLSA an appealing choice for latent response analysis such as ranking latent topics by their factorized response values. We apply the proposed sPLSA model to analyze the controversy of bills from the United States Congress. We demonstrate the effectiveness of our model by identifying contentious legislative issues. [ABSTRACT FROM AUTHOR]

Topic models provide a useful text-mining tool for learning, extracting, and discovering latent structures in large text corpora. Although a plethora of methods have been proposed for topic modeling, lacking in the literature is a formal theoretical investigation of the statistical identifiability and accuracy of latent topic estimation. In this article, we propose a maximum likelihood estimator (MLE) of latent topics based on a specific integrated likelihood that is naturally connected to the concept, in computational geometry, of volume minimization. Our theory introduces a new set of geometric conditions for topic model identifiability, conditions that are weaker than conventional separability conditions, which typically rely on the existence of pure topic documents or of anchor words. Weaker conditions allow a wider and thus potentially more fruitful investigation. We conduct finite-sample error analysis for the proposed estimator and discuss connections between our results and those of previous investigations. We conclude with empirical studies employing both simulated and real datasets. for this article are available online. [ABSTRACT FROM AUTHOR]

Abstract Cell-type composition is an important indicator of health. We present Guided Topic Model for deconvolution (GTM-decon) to automatically infer cell-type-specific gene topic distributions from single-cell RNA-seq data for deconvolving bulk transcriptomes. GTM-decon performs competitively on deconvolving simulated and real bulk data compared with the state-of-the-art methods. Moreover, as demonstrated in deconvolving disease transcriptomes, GTM-decon can infer multiple cell-type-specific gene topic distributions per cell type, which captures sub-cell-type variations. GTM-decon can also use phenotype labels from single-cell or bulk data to infer phenotype-specific gene distributions. In a nested-guided design, GTM-decon identified cell-type-specific differentially expressed genes from bulk breast cancer transcriptomes.

This article provides a structured description of openly available news topics and forecasts for armed conflict at the national and grid cell level starting January 2010. The news topics, as well as the forecasts, are updated monthly at conflictforecast.org and provide coverage for more than 170 countries and about 65,000 grid cells of size 55 × 55 km worldwide. The forecasts rely on natural language processing (NLP) and machine learning techniques to leverage a large corpus of newspaper text for predicting sudden onsets of violence in peaceful countries. Our goals are a) to support conflict prevention efforts by making our risk forecasts available to practitioners and research teams worldwide, b) to facilitate additional research that can utilize risk forecasts for causal identification, and c) to provide an overview of the news landscape.

The term “climate engineering” is an ambiguous label for two categories of technologies for mediating global warming, Carbon Dioxide Removal (CDR) and Solar Radiation Modification (SRM). Whether CDR and SRM should be grouped together under a common umbrella term, or whether they should be treated as two disparate problems in need of disparate solutions, has long been a matter of debate. This paper first provides an empirical analysis of disaggregation. Topic modelling the scientific literature on climate engineering, I highlight a trend towards disaggregation driven by growth in the specialized literature on CDR. Second, I explore inconsistencies in the theoretical rationale for disaggregation and challenges in its practical implications. Third, I elaborate on the theoretical and practical utility of maintaining an integrated concept of climate engineering for highlighting the challenges of governing large-scale, hypothetical technological interventions into the climate system subject to deep uncertainties and mitigation deterrence.

Online news outlets have the power to influence public policy issues. To understand the opinions of the people, many government departments check online news outlets to manually detect events that interest people. This process is time-consuming. To promptly respond to public expectations, this research proposes a framework for detecting news events that may interest government departments. This article proposes a method for finding event trigger words used to represent an event. The news media can be a critical participant in 'agenda-setting', which means that more widely discussed news is more attractive and critical than news that is less discussed. However, few studies have considered the influence of news media publishers from the 'agenda setting' perspective. Therefore, this study proposes an 'agenda setting'-based filter to establish a high-impact news event detection model. The proposed framework identifies trigger words and utilises word embedding to find news event–related words. After that, an event detection model is designed to determine the events that are attractive to government departments. The experimental results show that purity increases from 0.666 when no extraction method is used to 0.809 when the extraction method in this study is used. The overall improvement trend shows significant improvement in event detection performance. [ABSTRACT FROM AUTHOR]

There is a clear consensus among climate scientists about the reality and serious consequences of anthropogenic climate change. However, a vocal minority challenges this consensus. While some research has drawn attention to how conservative foundations support these anti-consensus scientists, less is known about how these scholars are embedded within the broader scientific community. Here, we analyze the networks of anti-consensus and consensus scientists and observe the extent to which these groups are maintained through peer collaborations (e.g. co-authorship) or substantive focus (e.g. research specialization). Using bibliometric data, we construct co-authorship and bibliographic networks linking scientists that appear in two key reports representing the consensus and anti-consensus positions. We identify specialty areas using text analysis and model participation in either series of reports. Results indicate that anti-consensus scientists are not in the same network as consensus scientists and have somewhat different research specializations than consensus scientists although there is substantive overlap. Additionally, anti-consensus scientists do not form a coherent network among themselves, which suggests they do not constitute a separate scientific community, but rather are composed of a disparate group of idiosyncratic scientists. • Clear consensus exists among climate scientists about anthropogenic climate change, yet a small minority is anti-consensus. • Consensus and anti-consensus scientists differ by collaboration and somewhat by specialization, among other factors. • Consensus scientists form a community, while anti-consensus scientists have few ties to consensus scientists or each other. [ABSTRACT FROM AUTHOR]

Since the early twentieth century, scholarly interest in the intersection between knowledge and political rationality in advanced liberal democracies has drawn attention to two general and contradictory processes: first, the rise of technocracy, and of the institutions, and experts, who use technical knowledge as a lever of power; second, the democratization of expertise – the emergence of lay audiences as stake-holders and competent participants in technical and scientific decisions and debates. In this paper, I analyze the annual reports on the Italian economy written by the Bank of Italy between 1960 and 1984, and trace the debate they spurred in three national newspaper outlets. I detail the emergence of public expertise on the economy, as well as the emergence of crisis and reputation management as techniques for the Bank to bolster its authority. I argue that the Bank of Italy, by framing the present as an exception, achieved a form of performativity that I call fictional. [ABSTRACT FROM AUTHOR]

The existing topic models usually focus the citation relationships when considering the relationships between documents. To make full use of the information contained in scientific documents, this paper proposes a Multi-Relational Neural Topic Model (MRNTM) based on the three relationships networks between documents. This work comprehensively considers the multiple relationships networks between documents: citation relationship, author relationship, and co-citation relationship, and then proposes a model that combines Variational Auto-Encoder (VAE) with neural networks multi-layer perceptron (MLP). This model not only can be used to learn more representative document topics, but also capture the complex interaction between documents according to their latent topics. The interaction between documents can further promote topic learning. Experiments on two real datasets show that our model can effectively utilize latent topics and the relationship between document networks, and superior to existing models in topic learning and citation link prediction. [ABSTRACT FROM AUTHOR]

In scientometrics, scientific collaboration is often analyzed by means of co-authorships. An aspect which is often overlooked and more difficult to quantify is the flow of expertise between authors from different research topics, which is an important part of scientific progress. With the Topic Flow Network (TFN) we propose a graph structure for the analysis of research topic flows between scientific authors and their respective research fields. Based on a multi-graph and a topic model, our proposed network structure accounts for intratopic as well as intertopic flows. Our method requires for the construction of a TFN solely a corpus of publications (i.e., author and abstract information). From this, research topics are discovered automatically through non-negative matrix factorization. The thereof derived TFN allows for the application of social network analysis techniques, such as common metrics and community detection. Most importantly, it allows for the analysis of intertopic flows on a large, macroscopic scale, i.e., between research topic, as well as on a microscopic scale, i.e., between certain sets of authors. We demonstrate the utility of TFNs by applying our method to two comprehensive corpora of altogether 20 Mio. publications spanning more than 60 years of research in the fields computer science and mathematics. Our results give evidence that Topic Flow Networks are suitable, e.g., for the analysis of topical communities, the discovery of important authors in different fields, and, most notably, the analysis of intertopic flows, i.e., the transfer of topical expertise. Besides that, our method opens new directions for future research, such as the investigation of influence relationships between research fields. [ABSTRACT FROM AUTHOR]

Gaining an understanding of the concerns and aspirations of a country's diaspora can help domestic politicians to better connect with this community and gain their support in elections. The United Kingdom's diaspora is large and spread among many countries, and currently has the right to vote in UK general elections only for a limited time. However, there are proposals to abolish these time limits and this will make this community of increasing interest to politicians. This study uses signatories to the UK Parliaments e-petitions platform to gain an understanding of the foreign and domestic political concerns of this community. The analysis uses Latent Dirichlet allocation to identify common topics among the e-petitions and hierarchical clustering to identify commonalities among countries, territories and regions. It is found that there are five meaningful groups of such, and they are diverse in the topics that are of most concern. [ABSTRACT FROM AUTHOR]

Marketing research relies on individual-level estimates to understand the rich heterogeneity of consumers, firms, and products. While much of the literature focuses on capturing static cross-sectional heterogeneity, little research has been done on modeling dynamic heterogeneity, or the heterogeneous evolution of individual-level model parameters. In this work, the authors propose a novel framework for capturing the dynamics of heterogeneity, using individual-level, latent, Bayesian nonparametric Gaussian processes. Similar to standard heterogeneity specifications, this Gaussian process dynamic heterogeneity (GPDH) specification models individual-level parameters as flexible variations around population-level trends, allowing for sharing of statistical information both across individuals and within individuals over time. This hierarchical structure provides precise individual-level insights regarding parameter dynamics. The authors show that GPDH nests existing heterogeneity specifications and that not flexibly capturing individual-level dynamics may result in biased parameter estimates. Substantively, they apply GPDH to understand preference dynamics and to model the evolution of online reviews. Across both applications, they find robust evidence of dynamic heterogeneity and illustrate GPDH's rich managerial insights, with implications for targeting, pricing, and market structure analysis. [ABSTRACT FROM AUTHOR]

This study focused on science communication on the websites of Czech research institutions. Particularly, we inquired to what extent Czech science is shared with the public on the Internet and what differences can be found between the websites of social and natural science institutions. Textual analysis revealed that on the scientific websites, terms like ‘science’ and ‘popularization’ occurred together with references to scientific institutions, study, and research. In the case of natural sciences, the term ‘popularization’ was more often linked to receiving science awards for science popularization and promotion. Structural web analysis showed that most scientific webs contained hyperlinks to social media such as Facebook, Twitter, YouTube, Instagram, and LinkedIn. Similarly, they often referred to online news outlets such as ceskatelevize.cz, novinky.cz, lidovky.cz, and rozhlas.cz. On the other side, they much less often referred to institutional and government websites. The results suggested that Czech science communication can be characterized as more interactive than canonical. [ABSTRACT FROM AUTHOR]

This study examines how emergent Twitter publics are organised and engage with political scandal and personalisation during Covid-19 in the UK. The analysis is centred on a series of media events around Chief Adviser to the then-UK Prime Minister, running from May 2020 to May 2021. The samples comprises original tweets that contain key hashtags, amounting to 38,326 items. These are subject to topic model analysis to identify semantic fields, before using critical discourse analysis. We find hashtags help constitute emergent Twitter publics, and that tweets follow conversational patterns and conspire in tactics of intertextuality. Dissention to government conduct engages resourcefully with the affordances of Twitter: constituting publics, shaping discourse, and articulating with parallel discussions on political performance. Further, a computational approach can systematise the identification of domains of discourse and relevant lexical sets, providing an evidence-based understanding of even novel and emergent political discourses in online discussion. [ABSTRACT FROM AUTHOR]

Topic models are a useful and popular method to find latent topics of documents. However, the short and sparse texts in social media micro-blogs such as Twitter are challenging for the most commonly used Latent Dirichlet Allocation (LDA) topic model. We compare the performance of the standard LDA topic model with the Gibbs Sampler Dirichlet Multinomial Model (GSDMM) and the Gamma Poisson Mixture Model (GPM), which are specifically designed for sparse data. To compare the performance of the three models, we propose the simulation of pseudo-documents as a novel evaluation method. In a case study with short and sparse text, the models are evaluated on tweets filtered by keywords relating to the Covid-19 pandemic. We find that standard coherence scores that are often used for the evaluation of topic models perform poorly as an evaluation metric. The results of our simulation-based approach suggest that the GSDMM and GPM topic models may generate better topics than the standard LDA model. [ABSTRACT FROM AUTHOR]

In this study, we propose a recognition method of measuring literature topic evolution paths based on K-means-NMF in order to address problems such as the unobvious effect of topic clustering, high degree of mixing in clustering results, and unclear topic evolution paths that exist in the current research of topic evolution analysis. Firstly, we enhance the traditional NMF (Nonnegative Matrix Factorization) topic model by combining the Kmeans clustering algorithm with the NMF model to improve the accuracy of topic clustering and reduce the correlation among topics. Secondly, we perform the topic co-occurrence analysis based on the clustering results to identify important topic categories for recognizing critical evolution paths to solve the problem of multiple possible evolution paths in the experiment. Thirdly, we adopt the Word2Vec model to calculate topic word vectors in a semantic context to improve the accuracy of the correlation strength between topics at adjacent stages. Finally, we adopt the above method to conduct an empirical study using intelligent algorithms as an example. The experimental results show that this research method effectively identifies important topics and topic developments in the subject area, which can support scientific research and science and technology policy development. [ABSTRACT FROM AUTHOR]

Recent calls from within economics for increased attention to narrative open the door to possible cross-fertilisation between economics and more humanistically oriented business and economic history. Indeed, arguments for economists to take narratives seriously and incorporate them into economic theory have some similarities with classic calls for a revival of narrative in history and abandonment of 'scientific' history. Both share an approach to explaining social phenomena based on the micro-level. This article examines how new methods in computational text analysis can be employed to further the goals of prioritising narrative in economics and history but also challenge a focus on the micro-level. Through a survey of the most frequently used tools of computational text analysis and an overview of their uses to date across the social sciences and humanities, this article shows how such methods can provide economic and business historians tools to respond to and engage with the 'narrative turn' in economics while also building on and offering a macro-level corrective to the focus on narrative in history. [ABSTRACT FROM AUTHOR]

Examining teachers' knowledge on a large scale involves addressing substantial measurement and logistical issues; thus, existing teacher knowledge assessments have mainly consisted of selected-response items because of their ease of scoring. Although open-ended responses could capture a more complex understanding of and provide further insights into teachers' thinking, scoring these responses is expensive and time consuming, which limits their use in large-scale studies. In this study, we investigated whether a novel statistical approach, topic modeling, could be used to score teachers' open-ended responses and if so, whether these scores would capture nuances of teachers' understanding. To test this hypothesis, we used topic modeling to analyze teachers' responses to a proportional reasoning task and examined the associations of the topics identified through this method with categories identified by a separate qualitative analysis of the same data as well as teachers' performance on a measure of ratios and proportional relationships. Our findings suggest that topic modeling seemed to capture nuances of teachers' responses and that such nuances differentiated teachers' performance on the same concept. We discuss the implications of this study for education research. [ABSTRACT FROM AUTHOR]