Improving Sentiment Prediction of Textual Tweets Using Feature Fusion and Deep Machine Ensemble Model

Widespread fear and panic has emerged about COVID-19 on social media platforms which are often supported by falsified and altered content. This mass hysteria creates public anxiety due to misinformation, misunderstandings, and ignorance of the impact of COVID-19. To assist health professionals in addressing this epidemic more appropriately at the onset, sentiment analysis can potentially help the authorities for devising appropriate strategies. This study analyzes tweets related to COVID-19 using a machine learning approach and offers a high-accuracy solution. Experiments are performed involving different machine and deep learning models along with various features such as Word2vec, term-frequency, term-frequency document frequency, and feature fusion of both feature-generating approaches. The proposed approach combines the extra tree classifier and convolutional neural network and uses feature fusion to achieve the highest accuracy score of 99%. The proposed approach obtains far better results than existing sentiment analysis approaches. © 2023 by the authors.

Prevalence of SARS-COV-2 and other respiratory pathogens among a Jordanian subpopulation during Delta-to-Omicron transition: Winter 2021/2022.

Acute respiratory tract infections (ARTIs) during the winter months are associated with higher morbidity and mortality compared to other seasons of the year, with children below five, elderly, and immunocompromised patients being the most susceptible. Influenza A and B viruses, rhinovirus, coronaviruses, respiratory syncytial virus, adenovirus, and parainfluenza viruses, are the most frequently identified causes of viral ARTIs. In addition, the emergence of SARS-CoV-2 in 2019 provided an additional viral cause of ARTIs. The aim of this study was to provide an overview of the epidemiological status of upper respiratory infections, their main causative agents, and reported clinical presentation in the winter months of 2021, during two important surges of COVID-19 in Jordan. Nasopharyngeal samples were collected from 339 symptomatic patients during the period from December 2021 to March 2022, followed by nucleic acid isolation using a Viral RNA/DNA extraction Kit. The causative virus species associated with the patient's respiratory symptoms was determined utilizing a multiplex real-time PCR targeting 21 viruses, 11 bacteria, and a single fungus. SARS-CoV-2 was identified in 39.2% of the patients (n = 133/339). A total of 15 different pathogens were also identified as co-infections among these 133 patients (n = 67/133). SARS-CoV-2-Bacterial coinfections (37.6%, n = 50/133) were the most frequent, with Bordetella species being the most common, followed by Staphylococcus aureus, and H.influenzae type B. Viral coinfection rate was 27.8% (n = 37/133), with Influenza B virus and Human bocavirus being the most common. In Conclusion, Both SARS-CoV-2, influenza B virus, and Bordetella accounted for the majority of infections in patients with URTI during the winter months of 2021-2022. Interestingly, more than 50% of the patients with symptoms of URTIs were confirmed to have a coinfection with two or more respiratory pathogens, with SARS-CoV-2 and Bordetella coinfection being most predominant.

Immunoglobulins response of COVID-19 patients, COVID-19 vaccine recipients, and random individuals.

BACKGROUND: The development of specific immunoglobulins to COVID-19 after natural infection or vaccination has been proposed. The efficacy and dynamics of this response are not clear yet. AIM: This study aims to analyze the immunoglobulins response among COVID-19 patients, COVID-19 vaccine recipients and random individuals. METHODS: A total of 665 participants including 233 COVID-19 patients, 288 COVID-19 vaccine recipients, and 144 random individuals were investigated for anti-COVID-19 immunoglobulins (IgA, IgG, IgM). RESULTS: Among COVID-19 patients, 22.7% had detectable IgA antibodies with a mean of 27.3±57.1 ng/ml, 29.6% had IgM antibodies with a mean of 188.4±666.0 BAU/ml, while 59.2% had IgG antibodies with a mean of 101.7±139.7 BAU/ml. Pfizer-BioNTech vaccine recipients had positive IgG in 99.3% with a mean of 515.5±1143.5 BAU/ml while 85.7% of Sinopharm vaccine recipients had positive IgG with a mean of 170.0±230.0 BAU/ml. Regarding random individuals, 54.9% had positive IgG with a mean of 164.3±214 BAU/ml. The peak IgM response in COVID-19 patients was detected early at 15-22 days, followed by IgG peak at 16-30 days, and IgA peak at 0-60 days. IgM antibodies disappeared at 61-90 days, while IgG and IgA antibodies decreased slowly after the peak and remained detectable up to 300 days. The frequency of IgG positivity among patients was significantly affected by increased age, admission department (inpatient or outpatient), symptoms, need for oxygen therapy, and increased duration between positive COVID-19 RT PCR test and serum sampling (p˂0.05). Positive correlations were noted between different types of immunoglobulins (IgG, IgM, and IgA) among patients. CONCLUSIONS: Natural infection and COIVD-19 vaccines provide IgG-mediated immunity. The class, positivity, mean, efficacy, and duration of immunoglobulins response are affected by the mechanism of immunity and host related variables. Random community individuals had detectable COVID-19 IgG at ~55%, far from reaching herd immunity levels.

Cloudlet Federation Based Context-Aware Federated Learning Approach

A Cloudlet federation can be beneficial to overcome the latency and resource scarcity challenges in a cloudlet deployment altogether, as a task can run on a cloudlet within the federation, sharing resources of member cloudlets. Nonetheless, the cloudlet federation is not context-aware in terms of latency, so to perform federated learning in cloudlet federation, the selection of a resource-efficient deep learning model is challenging. Additionally, the accuracy of a deep learning model can be affected if end-user devices are unreliable and provide incorrect data for training deep learning models at the cloudlets. Thus, resource and context-aware federated learning solutions are required for accurate and latency-critical applications such as COVID-19 detection using X-ray images. This paper presents a novel context-aware cloudlet federated learning solution for COVID-19 detection that monitors the resources of a cloudlet using a broker thereby minimizing latency without any impact on the accuracy of the deep learning model. Results show that the proposed model reduces the latency by 5% and increases the accuracy by 5% as compared to the state-of-the-art conventional federated learning approach.

Selective Feature Sets Based Fake News Detection for COVID-19 to Manage Infodemic

During the COVID-19 pandemic, the spread of fake news became easy due to the wide use of social media platforms. Considering the problematic consequences of fake news, efforts have been made for the timely detection of fake news using machine learning and deep learning models. Such works focus on model optimization and feature engineering and the extraction part is under-explored area. Therefore, the primary objective of this study is to investigate the impact of features to obtain high performance. For this purpose, this study analyzes the impact of different subset feature selection techniques on the performance of models for fake news detection. Principal component analysis and Chi-square are investigated for feature selection using machine learning and pre-trained deep learning models. Additionally, the influence of different preprocessing steps is also analyzed regarding fake news detection. Results obtained from comprehensive experiments reveal that the extra tree classifier outperforms with a 0.9474 accuracy when trained on the combination of term frequency-inverse document frequency and bag of words features. Models tend to yield poor results if no preprocessing or partial processing is carried out. Convolutional neural network, long short term memory network, residual neural network (ResNet), and InceptionV3 show marginally lower performance than the extra tree classifier. Results reveal that using subset features also helps to achieve robustness for machine learning models.

IMPACT OF THE ONLINE LEARNING TECHNIQUES ON STUDENTS' CUMULATIVE GRADE POINTS

When COVID-19 prevailed, the educational system was shifted to online rather than traditional to facilitate the learning process. This study aimed at exploring the impacts of online learning techniques on the students' Cumulative Grade Point Average (CGPA). A total of 155 randomly selected students currently studying M. Phill education at the University of Agriculture Faisalabad, Pakistan participated in this study. Data were collected through validated, pre-tested and reliable questionnaires. Collected data were analyzed using Statistical Package for Social Sciences (SPSS). Findings unveiled that online learning techniques improved the learning abilities, personality traits and teaching styles as perceived by the respondents which further improved the CGPA of students. Within the effects on learning abilities, enabling students to judicious use of technology, multimedia, observation and clearing the concepts were major improvements which helped students to attain an increase in CGPA. As for as effects on personality traits were concerned, social interaction enhanced communication skills and improvement in understanding, social skills and confidence led the students to get high CGPA. Moreover, online learning improved the teaching styles by integrating video lectures, immediate results assessment and easy access to the technology were key drivers of the increase in CGPA. This study suggested a hybrid educational system at the University of Agriculture Faisalabad for effective learning in students.

Outcomes of COVID-19 in hospitalized solid organ transplant recipients compared to a matched cohort of non-transplant patients at a national healthcare system in the United States.

Data describing outcomes of solid organ transplant (SOT) recipients with coronavirus disease 2019 (COVID-19) are variable, and the association between SOT status and mortality remains unclear. In this study, we compare clinical outcomes of SOT recipients hospitalized with COVID-19 between March 10, and September 1, 2020, to a matched cohort of non-SOT recipients at a national healthcare system in the United States (US). From a population of 43 461 hospitalized COVID-19-positive patients, we created a coarsened exact matched cohort of 4035 patients including 128 SOT recipients and 3907 weighted matched non-SOT controls. Multiple logistic regression was used to evaluate association between SOT status and clinical outcomes. Among the 4035 patients, median age was 60 years, 61.7% were male, 21.9% were Black/African American, and 50.8% identified as Hispanic/Latino ethnicity. Patients with a history of SOT were more likely to die within the study period when compared to matched non-SOT recipients (21.9% and 14.9%, respectively; odds ratio [OR] 1.93; 95% confidence interval [CI]: 1.18-3.15). Moreover, SOT status was associated with increased odds of receiving invasive mechanical ventilation (OR [95% CI]: 2.34 [1.51-3.65]), developing acute kidney injury (OR [95% CI]: 2.41 [1.59-3.65]), and receiving vasopressor support during hospitalization (OR [95% CI]: 2.14 [1.31-3.48]).

Could "trained immunity" be induced by live attenuated vaccines protect against COVID-19? Review of available evidence.

Coronavirus disease 2019 (COVID-19) represents a severe global public health threat. Caused by SARS-Cov-2, COVID-19 is characterized by high transmission rate that correlates with high viral load. The full clinical spectrum of the illness, the prevalence rates of mild symptomatic and asymptomatic cases, and the case fatality rates are still poorly understood, highlighting the importance of early preventive measures. Unfortunately, appropriate vaccination against SARS-Cov-2 is not yet available. Unless a target vaccine is developed, COVID-19 impacts will be devastating. "Trained immunity" (TI), which could be induced by live attenuated vaccines (LAVs), is a potential public health preventive approach to boost the host immune system. Trained innate immune cells demonstrated phenotypical and functional changes leading them to acquire immunological memory and amplify their responses against subsequent infections. This phenomenon could have important public health preventive implications by harnessing the early immune responses against COVID-19, restricting its progression, and suppressing its infectivity. Some LAVs have induced a broad, nonspecific, protection against unrelated pathogens and decreased mortality from conditions other than the targeted infectious diseases. This review summarizes the relevant literature and 1) emphasizes the role of available LAVs as potential stimulants for TI and 2) proposes this phenomenon as a potential preventive approach against COVID-19 that needs thoughtful consideration and further investigation. Clinical trials in this field are then urgently needed in line of vaccine and treatment unavailability. This is specifically true when considering two evolving scenarios; the virus spread may not diminish with warm weather, and that it will erupt a second-hit severe outbreak next winter.

Medical Students and COVID-19: Knowledge, Attitudes, and Precautionary Measures. A Descriptive Study From Jordan.

The recent coronavirus disease (COVID-19) pandemic is associated with increasing morbidity and mortality and has impacted the lives of the global populations. Human behavior and knowledge assessment during the crisis are critical in the overall efforts to contain the outbreak. To assess knowledge, attitude, perceptions, and precautionary measures toward COVID-19 among a sample of medical students in Jordan. This is a cross-sectional descriptive study conducted between the 16th and 19th of March 2020. Participants were students enrolled in different levels of study at the six medical schools in Jordan. An online questionnaire which was posted on online platforms was used. The questionnaire consisted of four main sections: socio-demographics, sources of information, knowledge attitudes, and precautionary measures regarding COVID-19. Medical students used mostly social media (83.4%) and online search engines (84.8%) as their preferred source of information on COVID-19 and relied less on medical search engines (64.1%). Most students believed that hand shaking (93.7%), kissing (94.7%), exposure to contaminated surfaces (97.4%), and droplet inhalation (91.0%) are the primary mode of transmission but were indecisive regarding airborne transmission with only 41.8% in support. Participants also reported that elderly with chronic illnesses are the most susceptible group for the coronavirus infection (95.0%). As a response to the COVID-19 pandemic more than 80.0% of study participants adopted social isolation strategies, regular hand washing, and enhanced personal hygiene measures as their first line of defense against the virus. In conclusion, Jordanian medical students showed expected level of knowledge about the COVID-19 virus and implemented proper strategies to prevent its spread.