Pseudo-document simulation for comparing LDA, GSDMM and GPM topic models on short and sparse text using Twitter data.

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.

Prevalence dependent calibration of a predictive model for nasal carriage of methicillin-resistant Staphylococcus aureus.

BACKGROUND: Published models predicting nasal colonization with Methicillin-resistant Staphylococcus aureus among hospital admissions predominantly focus on separation of carriers from non-carriers and are frequently evaluated using measures of discrimination. In contrast, accurate estimation of carriage probability, which may inform decisions regarding treatment and infection control, is rarely assessed. Furthermore, no published models adjust for MRSA prevalence. METHODS: Using logistic regression, a scoring system (values from 0 to 200) predicting nasal carriage of MRSA was created using a derivation cohort of 3091 individuals admitted to a European tertiary referral center between July 2007 and March 2008. The expected positive predictive value of a rapid diagnostic test (GeneOhm, Becton & Dickinson Co.) was modeled using non-linear regression according to score. Models were validated on a second cohort from the same hospital consisting of 2043 patients admitted between August 2008 and January 2012. Our suggested correction score for prevalence was proportional to the log-transformed odds ratio between cohorts. Calibration before and after correction, i.e. accurate classification into arbitrary strata, was assessed with the Hosmer-Lemeshow-Test. RESULTS: Treating culture as reference, the rapid diagnostic test had positive predictive values of 64.8% and 54.0% in derivation and internal validation corhorts with prevalences of 2.3% and 1.7%, respectively. In addition to low prevalence, low positive predictive values were due to high proportion (> 66%) of mecA-negative Staphylococcus aureus among false positive results. Age, nursing home residence, admission through the medical emergency department, and ICD-10-GM admission diagnoses starting with "A" or "J" were associated with MRSA carriage and were thus included in the scoring system, which showed good calibration in predicting probability of carriage and the rapid diagnostic test's expected positive predictive value. Calibration for both probability of carriage and expected positive predictive value in the internal validation cohort was improved by applying the correction score. CONCLUSIONS: Given a set of patient parameters, the presented models accurately predict a) probability of nasal carriage of MRSA and b) a rapid diagnostic test's expected positive predictive value. While the former can inform decisions regarding empiric antibiotic treatment and infection control, the latter can influence choice of screening method.

Unsupervised document classification integrating web scraping, one-class SVM and LDA topic modelling.

Unsupervised document classification for imbalanced data sets poses a major challenge. To obtain accurate classification results, training data sets are often created manually by humans which requires expert knowledge, time and money. Depending on the imbalance of the data set, this approach also either requires human labelling of all of the data or it fails to adequately recognize underrepresented categories. We propose an integration of web scraping, one-class Support Vector Machines (SVM) and Latent Dirichlet Allocation (LDA) topic modelling as a multi-step classification rule that circumvents manual labelling. Unsupervised one-class document classification with the integration of out-of-domain training data is achieved and >80% of the target data is correctly classified. The proposed method thus even outperforms common machine learning classifiers and is validated on multiple data sets.

Mapping ex ante risks of COVID-19 in Indonesia using a Bayesian geostatistical model on airport network data.

A rapid response to global infectious disease outbreaks is crucial to protect public health. Ex ante information on the spatial probability distribution of early infections can guide governments to better target protection efforts. We propose a two-stage statistical approach to spatially map the ex ante importation risk of COVID-19 and its uncertainty across Indonesia based on a minimal set of routinely available input data related to the Indonesian flight network, traffic and population data, and geographical information. In a first step, we use a generalised additive model to predict the ex ante COVID-19 risk for 78 domestic Indonesian airports based on data from a global model on the disease spread and covariates associated with Indonesian airport network flight data prior to the global COVID-19 outbreak. In a second step, we apply a Bayesian geostatistical model to propagate the estimated COVID-19 risk from the airports to all of Indonesia using freely available spatial covariates including traffic density, population and two spatial distance metrics. The results of our analysis are illustrated using exceedance probability surface maps, which provide policy-relevant information accounting for the uncertainty of the estimates on the location of areas at risk and those that might require further data collection.

An iterative topic model filtering framework for short and noisy user-generated data: analyzing conspiracy theories on twitter.

Conspiracy theories have seen a rise in popularity in recent years. Spreading quickly through social media, their disruptive effect can lead to a biased public view on policy decisions and events. We present a novel approach for LDA-pre-processing called Iterative Filtering to study such phenomena based on Twitter data. In combination with Hashtag Pooling as an additional pre-processing step, we are able to achieve a coherent framing of the discussion and topics of interest, despite of the inherent noisiness and sparseness of Twitter data. Our novel approach enables researchers to gain detailed insights into discourses of interest on Twitter, allowing them to identify tweets iteratively that are related to an investigated topic of interest. As an application, we study the dynamics of conspiracy-related topics on US Twitter during the last four months of 2020, which were dominated by the US-Presidential Elections and Covid-19. We monitor the public discourse in the USA with geo-spatial Twitter data to identify conspiracy-related contents by estimating Latent Dirichlet Allocation (LDA) Topic Models. We find that in this period, usual conspiracy-related topics played a marginal role in comparison with dominating topics, such as the US-Presidential Elections or the general discussions about Covid-19. The main conspiracy theories in this period were the ones linked to "Election Fraud" and the "Covid-19-hoax." Conspiracy-related keywords tended to appear together with Trump-related words and words related to his presidential campaign.

The Rhinobiome of Exacerbated Wheezers and Asthmatics: Insights From a German Pediatric Exacerbation Network.

Although the nose, as a gateway for organism-environment interactions, may have a key role in asthmatic exacerbation, the rhinobiome of exacerbated children with asthma was widely neglected to date. The aim of this study is to understand the microbiome, the microbial immunology, and the proteome of exacerbated children and adolescents with wheeze and asthma. Considering that a certain proportion of wheezers may show a progression to asthma, the comparison of both groups provides important information regarding clinical and phenotype stratification. Thus, deep nasopharyngeal swab specimens, nasal epithelial spheroid (NAEsp) cultures, and blood samples of acute exacerbated wheezers (WH), asthmatics (AB), and healthy controls (HC) were used for culture (n = 146), 16 S-rRNA gene amplicon sequencing (n = 64), and proteomic and cytokine analyses. Interestingly, Proteobacteria were over-represented in WH, whereas Firmicutes and Bacteroidetes were associated with AB. In contrast, Actinobacteria commonly colonized HCs. Moreover, Staphylococcaceae, Enterobacteriaceae, Burkholderiaceae, Xanthobacteraceae, and Sphingomonadaceae were significantly more abundant in AB compared to WH and HC. The α-diversity analyses demonstrated an increase of bacterial abundance levels in atopic AB and a decrease in WH samples. Microbiome profiles of atopic WH differed significantly from atopic AB, whereby atopic samples of WH were more homogeneous than those of non-atopic subjects. The NAEsp bacterial exposure experiments provided a disrupted epithelial cell integrity, a cytokine release, and cohort-specific proteomic differences especially for Moraxella catarrhalis cultures. This comprehensive dataset contributes to a deeper insight into the poorly understood plasticity of the nasal microbiota, and, in particular, may enforce our understanding in the pathogenesis of asthma exacerbation in childhood.

On the 100th anniversary of sterile catgut kuhn: Franz Kuhn (1866-1929) and the epistemology of catgut sterilization.

The long road to effective catgut sterilization began with the work of Lord Joseph Lister (1867) and did not end until 40 years later. At the end of the nineteenth century dozens of different techniques were used to "sterilize" catgut, by immersing the cord in a cold chemical solution, by exposing it to steam, or by a combination of the two techniques, yet none of these approaches offered the ultimate solution. One of the many physicians working on the catgut problem at that time was the German surgeon Franz Kuhn (1866-1929), best known as a pioneer of intubation anesthesia. This review offers a brief biographical sketch of Kuhn's life and career on the occasion of the centenary of Sterile Catgut Kuhn. The goal of the present study is to describe several landmarks in the development of the catgut sterilization method. To explain this process, two approaches are taken: first, an analysis to see whether the character traits of the typical surgeon at that time provided the soil in which innovation could thrive, and second, an epistemological examination of the conceptual models for the attainment of knowledge current at that time. Perspectives for the future are explored in light of the "imperative of responsibility" of Hans Jonas.