COVID-19 and Media datasets: Period- and location-specific textual data mining.
Ontology highlight
ABSTRACT:
SUBMITTER: Roche M
PROVIDER: S-EPMC7525263 | biostudies-literature | 2020 Dec
REPOSITORIES: biostudies-literature
Publications
COVID-19 and Media datasets: Period- and location-specific textual data mining.
Data in brief 20200930
The vocabulary used in news on a disease such as COVID-19 changes according the period [4]. This aspect is discussed on the basis of MEDISYS-sourced media datasets via two studies. The first focuses on terminology extraction and the second on period prediction according to the textual content using machine learning approaches. ...[more]
Similar Datasets
Project description:The unpredictability and unknowns surrounding the ongoing coronavirus disease (COVID-19) pandemic have led to an unprecedented consequence taking a heavy toll on the lives and economies of all countries. There have been efforts to predict COVID-19 case counts (CCC) using epidemiological data and numerical tokens online, which may allow early preventive measures to slow the spread of the disease. In this paper, we use state-of-the-art natural language processing (NLP) algorithms to numerically encode COVID-19 related tweets originated from eight cities in the United States and predict city-specific CCC up to eight days in the future. A city-embedding is proposed to obtain a time series representation of daily tweets posted from a city, which is then used to predict case counts using a custom long-short term memory (LSTM) model. The universal sentence encoder yields the best normalized root mean squared error (NRMSE) 0.090 (0.039), averaged across all cities in predicting CCC six days in the future. The R 2 scores in predicting CCC are more than 0.70 and often over 0.8, which suggests a strong correlation between the actual and our model predicted CCC values. Our analyses show that the NRMSE and R 2 scores are consistently robust across different cities and different numbers of time steps in time series data. Results show that the LSTM model can learn the mapping between the NLP-encoded tweet semantics and the case counts, which infers that social media text can be directly mined to identify the future course of the pandemic.
| S-EPMC9490453 | biostudies-literature
Project description:By 29 May 2020, the coronavirus disease (COVID-19) caused by SARS-CoV-2 had spread to 188 countries, infecting more than 5.9 million people, and causing 361,249 deaths. Governments issued travel restrictions, gatherings of institutions were cancelled, and citizens were ordered to socially distance themselves in an effort to limit the spread of the virus. Fear of being infected by the virus and panic over job losses and missed education opportunities have increased people's stress levels. Psychological studies using traditional surveys are time-consuming and contain cognitive and sampling biases, and therefore cannot be used to build large datasets for a real-time depression analysis. In this article, we propose a CorExQ9 algorithm that integrates a Correlation Explanation (CorEx) learning algorithm and clinical Patient Health Questionnaire (PHQ) lexicon to detect COVID-19 related stress symptoms at a spatiotemporal scale in the United States. The proposed algorithm overcomes the common limitations of traditional topic detection models and minimizes the ambiguity that is caused by human interventions in social media data mining. The results show a strong correlation between stress symptoms and the number of increased COVID-19 cases for major U.S. cities such as Chicago, San Francisco, Seattle, New York, and Miami. The results also show that people's risk perception is sensitive to the release of COVID-19 related public news and media messages. Between January and March, fear of infection and unpredictability of the virus caused widespread panic and people began stockpiling supplies, but later in April, concerns shifted as financial worries in western and eastern coastal areas of the U.S. left people uncertain of the long-term effects of COVID-19 on their lives.
| S-EPMC7400345 | biostudies-literature
Project description:MotivationA global medical crisis like the coronavirus disease 2019 (COVID-19) pandemic requires interdisciplinary and highly collaborative research from all over the world. One of the key challenges for collaborative research is a lack of interoperability among various heterogeneous data sources. Interoperability, standardization and mapping of datasets are necessary for data analysis and applications in advanced algorithms such as developing personalized risk prediction modeling.ResultsTo ensure the interoperability and compatibility among COVID-19 datasets, we present here a common data model (CDM) which has been built from 11 different COVID-19 datasets from various geographical locations. The current version of the CDM holds 4639 data variables related to COVID-19 such as basic patient information (age, biological sex and diagnosis) as well as disease-specific data variables, for example, Anosmia and Dyspnea. Each of the data variables in the data model is associated with specific data types, variable mappings, value ranges, data units and data encodings that could be used for standardizing any dataset. Moreover, the compatibility with established data standards like OMOP and FHIR makes the CDM a well-designed CDM for COVID-19 data interoperability.Availability and implementationThe CDM is available in a public repo here: https://github.com/Fraunhofer-SCAI-Applied-Semantics/COVID-19-Global-Model.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC9750115 | biostudies-literature
Project description:In the epidemiological COVID-19 research, artificial intelligence is a unique approach to make predictions about disease severity to manage COVID-19 patients. A limitation of artificial intelligence is, however, the high risk of bias. We investigated the skill of data mining and machine learning, two advanced forms of artificial intelligence, to predict severe COVID-19 pneumonia based on routine laboratory tests. A sample of 4009 COVID-19 patients was divided into Severe (PaO2< 60 mmHg, 489 cases) and Non-Severe (PaO2 ≥ 60 mmHg, 3520 cases) groups according to blood hypoxemia on admission and their laboratory datasets analyzed by the R software and WEKA workbench. After curation, data were processed for the selection of the most influential features including hemogram, pCO2, blood acid-base balance, prothrombin time, inflammation biomarkers, and glucose. The best fit of variables was successfully confirmed by either the Multilayer Perceptron, a feedforward neural network algorithm that performed machine recognition of severe COVID-19 with 96.5% precision, or by the C4.5 software, a supervised learning algorithm based on an objective-predefined variable (severity) that generated a decision tree with 89.4% precision. Finally, a complex bivariate Pearson's correlation matrix combined with advanced hierarchical clustering (dendrograms) were conducted for knowledge discovery. The hidden structure of the datasets revealed shift patterns related to the development of COVID-19-induced pneumonia that involved the lymphocyte-to-C-reactive protein and leukocyte-to-C-protein ratios, neutrophil %, pH and pCO2. The data mining approaches to the hematological fluctuations associated with severe COVID-19 pneumonia could not only anticipate adverse clinical outcomes, but also reveal putative therapeutic targets.
| S-EPMC8349478 | biostudies-literature
Project description:In this article we want to show the potential of an evolutionary algorithm called Topological Weighted Centroid (TWC). This algorithm can obtain new and relevant information from extremely limited and poor datasets. In a world dominated by the concept of big (fat?) data we want to show that it is possible, by necessity or choice, to work profitably even on small data. This peculiarity of the algorithm means that even in the early stages of an epidemic process, when the data are too few to have sufficient statistics, it is possible to obtain important information. To prove our theory, we addressed one of the most central issues at the moment: the COVID-19 epidemic. In particular, the cases recorded in Italy have been selected. Italy seems to have a central role in this epidemic because of the high number of measured infections. Through this innovative artificial intelligence algorithm, we have tried to analyze the evolution of the phenomenon and to predict its future steps using a dataset that contained only geospatial coordinates (longitude and latitude) of the first recorded cases. Once the coordinates of the places where at least one case of contagion had been officially diagnosed until February 26th, 2020 had been collected, research and analysis was carried out on: outbreak point and related heat map (TWC alpha); probability distribution of the contagion on February 26th (TWC beta); possible spread of the phenomenon in the immediate future and then in the future of the future (TWC gamma and TWC theta); how this passage occurred in terms of paths and mutual influence (Theta paths and Markov Machine). Finally, a heat map of the possible situation towards the end of the epidemic in terms of infectiousness of the areas was drawn up. The analyses with TWC confirm the assumptions made at the beginning.
| S-EPMC7382358 | biostudies-literature
Project description:BACKGROUND:We present an effective, rapid, systematic data mining approach for identifying genes or proteins related to a particular interest. A selected combination of programs exploring PubMed abstracts, universal gene/protein databases (UniProt, InterPro, NCBI Entrez), and state-of-the-art pathway knowledge bases (LSGraph and Ingenuity Pathway Analysis) was assembled to distinguish enzymes with hydrolytic activities that are expressed in the extracellular space of cancer cells. Proteins were identified with respect to six types of cancer occurring in the prostate, breast, lung, colon, ovary, and pancreas. RESULTS:The data mining method identified previously undetected targets. Our combined strategy applied to each cancer type identified a minimum of 375 proteins expressed within the extracellular space and/or attached to the plasma membrane. The method led to the recognition of human cancer-related hydrolases (on average, approximately 35 per cancer type), among which were prostatic acid phosphatase, prostate-specific antigen, and sulfatase 1. CONCLUSION:The combined data mining of several databases overcame many of the limitations of querying a single database and enabled the facile identification of gene products. In the case of cancer-related targets, it produced a list of putative extracellular, hydrolytic enzymes that merit additional study as candidates for cancer radioimaging and radiotherapy. The proposed data mining strategy is of a general nature and can be applied to other biological databases for understanding biological functions and diseases.
| S-EPMC1555615 | biostudies-literature
Project description:BackgroundThe high prevalence of COVID-19 has made it a new pandemic. Predicting both its prevalence and incidence throughout the world is crucial to help health professionals make key decisions. In this study, we aim to predict the incidence of COVID-19 within a two-week period to better manage the disease.MethodsThe COVID-19 datasets provided by Johns Hopkins University, contain information on COVID-19 cases in different geographic regions since January 22, 2020 and are updated daily. Data from 252 such regions were analyzed as of March 29, 2020, with 17,136 records and 4 variables, namely latitude, longitude, date, and records. In order to design the incidence pattern for each geographic region, the information was utilized on the region and its neighboring areas gathered 2 weeks prior to the designing. Then, a model was developed to predict the incidence rate for the coming 2 weeks via a Least-Square Boosting Classification algorithm.ResultsThe model was presented for three groups based on the incidence rate: less than 200, between 200 and 1000, and above 1000. The mean absolute error of model evaluation were 4.71, 8.54, and 6.13%, respectively. Also, comparing the forecast results with the actual values in the period in question showed that the proposed model predicted the number of globally confirmed cases of COVID-19 with a very high accuracy of 98.45%.ConclusionUsing data from different geographical regions within a country and discovering the pattern of prevalence in a region and its neighboring areas, our boosting-based model was able to accurately predict the incidence of COVID-19 within a two-week period.
| S-EPMC8182740 | biostudies-literature
Project description:Extracellular matrix (ECM) proteins have been shown to play important roles regulating multiple biological processes in an array of organ systems, including the cardiovascular system. Using a novel bioinformatics text-mining tool, we studied six categories of cardiovascular disease (CVD), namely, ischemic heart disease, cardiomyopathies, cerebrovascular accident, congenital heart disease, arrhythmias, and valve disease, anticipating novel ECM protein-disease and protein-protein relationships hidden within vast quantities of textual data. We conducted a phrase-mining analysis, delineating the relationships of 709 ECM proteins with the 6 groups of CVDs reported in 1,099,254 abstracts. The technology pipeline known as Context-Aware Semantic Online Analytical Processing was applied to semantically rank the association of proteins to each CVD and all six CVDs, performing analyses to quantify each protein-disease relationship. We performed principal component analysis and hierarchical clustering of the data, where each protein was visualized as a six-dimensional vector. We found that ECM proteins display variable degrees of association with the six CVDs; certain CVDs share groups of associated proteins, whereas others have divergent protein associations. We identified 82 ECM proteins sharing associations with all 6 CVDs. Our bioinformatics analysis ascribed distinct ECM pathways (via Reactome) from this subset of proteins, namely, insulin-like growth factor regulation and interleukin-4 and interleukin-13 signaling, suggesting their contribution to the pathogenesis of all six CVDs. Finally, we performed hierarchical clustering analysis and identified protein clusters predominantly associated with a targeted CVD; analyses of these proteins revealed unexpected insights underlying the key ECM-related molecular pathogenesis of each CVD, including virus assembly and release in arrhythmias. NEW & NOTEWORTHY The present study is the first application of a text-mining algorithm to characterize the relationships of 709 extracellular matrix-related proteins with 6 categories of cardiovascular disease described in 1,099,254 abstracts. Our analysis informed unexpected extracellular matrix functions, pathways, and molecular relationships implicated in the six cardiovascular diseases.
| S-EPMC6230912 | biostudies-literature
Project description:BackgroundIn December 2019, pneumonia cases of unknown origin were reported in Wuhan City, Hubei Province, China. Identified as the coronavirus disease (COVID-19), the number of cases grew rapidly by human-to-human transmission in Wuhan. Social media, especially Sina Weibo (a major Chinese microblogging social media site), has become an important platform for the public to obtain information and seek help.ObjectiveThis study aims to analyze the characteristics of suspected or laboratory-confirmed COVID-19 patients who asked for help on Sina Weibo.MethodsWe conducted data mining on Sina Weibo and extracted the data of 485 patients who presented with clinical symptoms and imaging descriptions of suspected or laboratory-confirmed cases of COVID-19. In total, 9878 posts seeking help on Sina Weibo from February 3 to 20, 2020 were analyzed. We used a descriptive research methodology to describe the distribution and other epidemiological characteristics of patients with suspected or laboratory-confirmed SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) infection. The distance between patients' home and the nearest designated hospital was calculated using the geographic information system ArcGIS.ResultsAll patients included in this study who sought help on Sina Weibo lived in Wuhan, with a median age of 63.0 years (IQR 55.0-71.0). Fever (408/485, 84.12%) was the most common symptom. Ground-glass opacity (237/314, 75.48%) was the most common pattern on chest computed tomography; 39.67% (167/421) of families had suspected and/or laboratory-confirmed family members; 36.58% (154/421) of families had 1 or 2 suspected and/or laboratory-confirmed members; and 70.52% (232/329) of patients needed to rely on their relatives for help. The median time from illness onset to real-time reverse transcription-polymerase chain reaction (RT-PCR) testing was 8 days (IQR 5.0-10.0), and the median time from illness onset to online help was 10 days (IQR 6.0-12.0). Of 481 patients, 32.22% (n=155) lived more than 3 kilometers away from the nearest designated hospital.ConclusionsOur findings show that patients seeking help on Sina Weibo lived in Wuhan and most were elderly. Most patients had fever symptoms, and ground-glass opacities were noted in chest computed tomography. The onset of the disease was characterized by family clustering and most families lived far from the designated hospital. Therefore, we recommend the following: (1) the most stringent centralized medical observation measures should be taken to avoid transmission in family clusters; and (2) social media can help these patients get early attention during Wuhan's lockdown. These findings can help the government and the health department identify high-risk patients and accelerate emergency responses following public demands for help.
| S-EPMC7236610 | biostudies-literature
Project description:BackgroundThe coronavirus disease (COVID-19) pandemic is considered to be the most daunting public health challenge in decades. With no effective treatments and with time needed to develop a vaccine, alternative approaches are being used to control this pandemic.ObjectiveThe objective of this paper was to identify topics, opinions, and recommendations about the COVID-19 pandemic discussed by medical professionals on the Twitter social medial platform.MethodsUsing a mixed methods approach blending the capabilities of social media analytics and qualitative analysis, we analyzed COVID-19-related tweets posted by medical professionals and examined their content. We used qualitative analysis to explore the collected data to identify relevant tweets and uncover important concepts about the pandemic using qualitative coding. Unsupervised and supervised machine learning techniques and text analysis were used to identify topics and opinions.ResultsData were collected from 119 medical professionals on Twitter about the coronavirus pandemic. A total of 10,096 English tweets were collected from the identified medical professionals between December 1, 2019 and April 1, 2020. We identified eight topics, namely actions and recommendations, fighting misinformation, information and knowledge, the health care system, symptoms and illness, immunity, testing, and infection and transmission. The tweets mainly focused on needed actions and recommendations (2827/10,096, 28%) to control the pandemic. Many tweets warned about misleading information (2019/10,096, 20%) that could lead to infection of more people with the virus. Other tweets discussed general knowledge and information (911/10,096, 9%) about the virus as well as concerns about the health care systems and workers (909/10,096, 9%). The remaining tweets discussed information about symptoms associated with COVID-19 (810/10,096, 8%), immunity (707/10,096, 7%), testing (605/10,096, 6%), and virus infection and transmission (503/10,096, 5%).ConclusionsOur findings indicate that Twitter and social media platforms can help identify important and useful knowledge shared by medical professionals during a pandemic.
| S-EPMC7304257 | biostudies-literature