Propaganda Identification on Twitter Platform During COVID-19 Pandemic Using LSTM

Online social media has been evolved as a universal platform for sharing information. Termination being shared on these platforms can be dubious or filthy. Propaganda is one of the systematic methods by which behavior of user can be manipulated. In this work, various machine learning methods are used for detecting such types of information on online social media. Data is collected d from Twitter using its API with the help of various ambiguous hashtags. The results showed that proposed Long Short Term Memory (LSTM) based propaganda identification showed better results than other machine learning techniques. An accuracy of 77.15% is achieved using the proposed approach. In the future BERT model can be used for achieving better Accuracy.

Weighted ensemble model for image classification.

The Deep Convolutional Neural Network (DCNN) classification models are being tremendously used across many research fields including medical science for image classification. The accuracy of the model and reliability on the results of the model are the key attributes which determine whether a particular model should be used for a specific application or not. A highly accurate model is always desirable for all applications of machine learning as well as deep learning. This paper presents a DCNN based heterogeneous ensemble approach where all DCNN models can be trained on a single dataset and each model can contribute of towards the final output of the ensemble model. The contribution of each model is weighted according to its individual accuracy on the given dataset. Models with higher accuracy has higher contribution in the final output of ensemble model, whereas the models with lower accuracy has lower contribution. This approach, when tested on two different X-ray images datasets of Covid-19, has confirmed the significant increase in 3-class accuracy as compared to the models in literature.

Hybrid Approach for Detecting Propagandistic Community and Core Node on Social Networks

People share their views and daily life experiences on social networks and form a network structure. The information shared on social networks can be unreliable, and detecting such kinds of information may reduce mass panic. Propaganda is a kind of biased or unreliable information that can mislead or intend to promote a political cause. The disseminators involved in spreading such information create a sophisticated network structure. Detecting such communities can lead to a safe and reliable network for the users. In this paper, a Boundary-based Community Detection Approach (BCDA) has been proposed to identify the core nodes in a propagandistic community that detects propagandistic communities from social networks with the help of interior and boundary nodes. The approach consists of two phases, one is to detect the community, and the other is to detect the core member. The approach mines nodes from the boundary as well as from the interior of the community structure. The leader Ranker algorithm is used for mining candidate nodes within the boundary, and the Constraint coefficient is used for mining nodes within the boundary. A novel dataset is generated from Twitter. About six propagandistic communities are detected. The core members of the propagandistic community are a combination of a few nodes. The experiments are conducted on a newly collected Twitter dataset consisting of 16 attributes. From the experimental results, it is clear that the proposed model outperformed other related approaches, including Greedy Approach, Improved Community-based 316 Robust Influence Maximization (ICRIM), Community Based Influence Maximization Approach (CBIMA), etc. It was also observed from the experiments that most of the propagandistic information is being shared during trending events around the globe, for example, at times of the COVID-19 pandemic.

Weighted ensemble model for image classification

The Deep Convolutional Neural Network (DCNN) classification models are being tremendously used across many research fields including medical science for image classification. The accuracy of the model and reliability on the results of the model are the key attributes which determine whether a particular model should be used for a specific application or not. A highly accurate model is always desirable for all applications of machine learning as well as deep learning. This paper presents a DCNN based heterogeneous ensemble approach where all DCNN models can be trained on a single dataset and each model can contribute of towards the final output of the ensemble model. The contribution of each model is weighted according to its individual accuracy on the given dataset. Models with higher accuracy has higher contribution in the final output of ensemble model, whereas the models with lower accuracy has lower contribution. This approach, when tested on two different X-ray images datasets of Covid-19, has confirmed the significant increase in 3-class accuracy as compared to the models in literature.

A Deep Neural Network for the Detection of COVID-19 from Chest X-ray Images

The most successful machine learning technology considered for analyzing a significant amount of chest X-ray images is Deep Learning and it has the potential to cause significant influence on Covid-19 screening. In this paper, we analyze four distinct Convolutional Neural Network (CNN) state-of-the-art architectures that are Baseline Model, Vanilla CNN, VGG-16 and Siamese Model on the basis of test accuracy. The effectiveness of the models under consideration is assessed using the Chest Radiograph dataset, which is publicly available for research. In order to discover COVID-19, we used well-known deep learning algorithms for data rarity. These include employing Siamese networks using transfer learning and a few-shot learning approach. Our experiments show that using few-shot learning methodologies, we can create a COVID-19 identification model that is both efficient and effective even with limited data. With this strategy, we were able to achieve 95% accuracy, compared to 86% with Baseline model. © 2022 Bharati Vidyapeeth, New Delhi.

COVID-19 and Pneumonia Detection using Deep Weighted Ensemble Model

The COVID-19 pandemic has brought us face to face with an unprecedented global crisis. However, it has also made us realize that problems like these can only be triumphed by the collectivization of human efforts and resources across the globe. Efforts towards achieving a higher accuracy in detection and classification of COVID-19 are of seminal importance, and this work is intended to contribute in that direction. The recent advances in Deep Learning, and its success, have proven to provide best and accurate solutions to medical imaging. In this study, we propose a Deep Ensemble model (Weighted Ensemble Model), in which we combine the output of five different state-of-art Deep Convolutional Neural Networks (DCNNs) wherein the output of each model is given a weight proportional to its individual accuracy on the given dataset. Weighted sum of the output of ensemble model is used to determine the class of input X-Ray image COVID-19, Pneumonia, Normal. Our approach of aggregating the outputs of various models helped us in achieving 100% accuracy, 100% precision and 100% recall as well. © 2022 Bharati Vidyapeeth, New Delhi.

Primary ureteroscopy for diagnosing and treating acute urolithiasis during the COVID-19 pandemic: Quality and cost benefits

Purpose: The purpose of this study was to investigate the management of acute urolithiasis during index admission by primary ureteroscopy (P-URS) during coronavirus disease-2019 (COVID-19) pandemic. With the rise in prevalence of urolithiasis, the focus has shifted to manage patients presenting with acute ureteric colic during their first admission rather than using temporary measures such as emergency stenting (ES) or nephrostomies which are followed by deferred ureteroscopic procedures Deferred Ureteroscopy (D-URS). We compared the results of ES with P-URS procedures in terms of quality and cost benefits during COVID-19 pandemic. Materials and Methods: Data were collected prospectively from April 2020 to March 2021 for all emergency urolithiasis procedures performed including ES and P-URS. The quality assessment was based in relation to patient factors including the number of procedures per patient, number of days spent at hospital, number of days off work, and expertise of person operating. Cost analysis included theater expenses, hospital stay charges, and loss of working days. Results: This study revealed that the average stay of patients on index admission who had an ES was 1.35 days compared to 1.78 days in patients who underwent P-URS. Patients who had ES had to undergo D-URS and spent another average of 1.5 days in the hospital. Overall, additional expenditure in patients who did not undergo primary ureterorenoscopy was on an average in the range of £1800 (excluding loss of work for patients, who needed to return for multiple procedures). Conclusion: We conclude that the approach of P-URS and management of stones in index admission is very effective in both improving quality of patients (during the COVID-19 pandemic) and bringing down cost expenditure effectively.

Clinical Characteristics and Outcomes in Patients Admitted in ICU with Severe COVID-19 in a Tertiary Care Hospital in North India: A Retrospective Observational Study

Background: Many studies have reported poor clinical outcomes regarding the ICU course of patients with severe COVID 19. Our study aimed at prospectively observing the predominant clinical pattern and outcomes in patients with severe COVID 19 admitted in the ICU. Materials and Methods: This study was a retrospective, observational study of 100 patients admitted to the ICU with confirmed COVID 19. Data from all patients with confirmed COVID 19 admitted in ICU between 15 March 2021 to 25 April 2021 was included for this study. Patients were studied for their demographics, baseline comorbidities, laboratory investigations, and details of treatment. Major outcomes analyzed were clinical presentation, mechanical ventilation (MV) related mortality, and overall mortality of ICU patients. Student's independent t-test for comparing continuous variables and Chi-square test for categorical variables. Results: Out of 220 patients with COVID-19, 100 were admitted to the ICU. The most common comorbidities were hypertension (38) and diabetes (25). Twenty-eight patients required mechanical ventilation (MV), out of which only 16 survived. MV LOS was longer for survivors than non-survivors. The overall mortality rate in ICU patients was 25%, and MV-related mortality was 42.85%. Conclusion: The severity of presenting symptoms and presentation time play a major role in the outcome. Our study reports higher mortality in patients who required mechanical ventilation. This could be because of the increased severity of disease in these patients. © 2021 Universitas Gadjah Mada - Faculty of Pharmacy. All rights reserved.

Delayed fixation of hip fractures and short-term outcome in coronavirus positive patients: A prospective cohort study

Introduction: Coronavirus Disease-19 (COVID-19) affected the health care system worldwide. The golden rules of fracture fixation and early mobilisation of patients was not strictly followed, because of fear of spread of the disease among the patients and health care workers. Early surgery and prompt postoperative ambulation improves outcomes for patients with hip fractures, but the morbidity and mortality were high in the patients who were operated upon, when having an active infection of COVID-19 virus. Aim: To study the short-term outcome of delayed fixation of hip fractures in coronavirus positive patient in terms of postoperative infection, union at the fracture, deep vein thrombosis and mortality. Materials and Methods: This was a prospective cohort study carried out at a tertiary care center in Kashmir, India from April 2020 to September 2020. Delayed surgery using different methods of fixation was performed in patients with hip fractures who had active COVID-19 infection. The patients were followed for a period of six months. Short term mortality and complications if any were recorded. Results were expressed in terms of frequency and percentages and analysed by Microsoft Excel 2016. Results: Among the 24 operated patients, males were 9 (37.5%) and females were 15 (62.5%). Of the total, 14 (58.33%) were intertrochanteric fractures, 6 (25%) were femoral neck and 4 (16.66%) were subtrochanteric fractures. Dynamic hip screw was used to treat 15 (62.5%) patients. Age ranged from 39 to 82 years mean age was 51.04 years. Majority of patients, 16 (66.66%) sustained hip fractures after a low velocity fall from standing height. Delay in surgery was 15 to 21 days (Average-18.25 days). Two elderly patients died after 14 weeks of follow-up due to causes else than respiratory failure. Three patients developed superficial infection which settled with oral antibiotics. No case of deep venous thrombosis, pulmonary thrombo embolism was observed in the present study. Conclusion: Despite the delay, the mortality rate in the early postoperative period was less. The present study findings suggest that hip fracture patients who present with COVID-19 infection can safely undergo delayed surgical intervention after appropriate medical optimisation.

Virtual Examination for Final Year Orthopaedics Postgraduate Residents during COVID-19: Is it a Viable Alternative?

Introduction: Final year examinations for postgraduate residents of medical colleges in India were scheduled according to a defined protocol where doctor patient relationship was given utmost importance. Due to the currently prevailing pandemic, it has become extremely unsafe to conduct such an examination. So, an innovative method was devised by the Department of Orthopaedics, GMC Srinagar to ensure safety of the examiners and candidates as well as maintaining the required standard of the examination. This method was employed successfully in conducting exit examinations of final year postgraduate residents in this department. Aim: To evaluate the effectiveness by a preformed questionnaire (developed by the faculty of Department of Orthopaedics, GMC Srinagar), Jammu and Kashmir, India of virtual pattern for exit examinations of final year postgraduate residents of medical colleges in India. Materials and Methods: A total number of 10 candidates appeared in the final year (MS Orthopaedics) examination. The conventional format which consisted of assigning patients to the candidates was replaced by a digital presentation format. In this format, the cases were prepared by the faculty, in the form of individual digital presentations for long as well as short cases. Presentation format consisted of detailed history, clinical examination, photographs of any deformity, elicited clinical signs and radiological imaging. A mock test was conducted four weeks before the date of the exam so that the candidates were familiar with the new virtual pattern. A questionnaire was designed to assess the acceptability of the virtual examination. It comprised of 10 questions and each answer was graded on a three point Likert point scale, producing a maximum score of 2 and a minimum score of 0 for each question. Results: Mean total scores for both the groups (examiners and candidates) were 18.25 and 16.9, respectively. The overall outcome regarding the responses was satisfactory in both the groups. The scoring was highest for the safety of the examination in examiner as well as the candidate group. In addition, overall satisfaction also scored the highest among the examiner group. Conclusion: This virtual pattern of examination presents a viable interim alternative to the traditional face to face examination, though it may not replace the latter. The time frame of the pandemic and its trend is difficult to imagine at this point of time, so, the virtual pattern might have to be used for future examinations as well. In addition, this pattern may even be utilised by other departments to frame their examinations.

Impact of unreliable content on social media users during COVID-19 and stance detection system

The abundant dissemination of misinformation regarding coronavirus disease 2019 (COVID-19) presents another unprecedented issue to the world, along with the health crisis. Online social network (OSN) platforms intensify this problem by allowing their users to easily distort and fabri-cate the information and disseminate it farther and rapidly. In this paper, we study the impact of misinformation associated with a religious inflection on the psychology and behavior of the OSN users. The article presents a detailed study to understand the reaction of social media users when exposed to unverified content related to the Islamic community during the COVID-19 lockdown period in India. The analysis was carried out on Twitter users where the data were collected using three scraping packages, Tweepy, Selenium, and Beautiful Soup, to cover more users affected by this misinformation. A labeled dataset is prepared where each tweet is assigned one of the four reaction polarities, namely, E (endorse), D (deny), Q (question), and N (neutral). Analysis of collected data was carried out in five phases where we investigate the engagement of E, D, Q, and N users, tone of the tweets, and the consequence upon repeated exposure of such information. The evidence demonstrates that the circulation of such content during the pandemic and lockdown phase had made people more vulnerable in perceiving the unreliable tweets as fact. It was also observed that people absorbed the negativity of the online content, which induced a feeling of hatred, anger, distress, and fear among them. People with similar mindset form online groups and express their negative attitude to other groups based on their opinions, indicating the strong signals of social unrest and public tensions in society. The paper also presents a deep learning-based stance detection model as one of the automated mechanisms for tracking the news on Twitter as being potentially false. Stance classifier aims to predict the attitude of a tweet towards a news headline and thereby assists in determining the veracity of news by monitoring the distribution of different reactions of the users towards it. The proposed model, employing deep learning (convolutional neural network (CNN)) and sentence embedding (bidirectional encoder representations from transformers (BERT)) techniques, outperforms the existing systems. The performance is evaluated on the benchmark SemEval stance dataset. Furthermore, a newly annotated dataset is prepared and released with this study to help the research of this domain. © 2020 by the authors. Licensee MDPI, Basel, Switzerland.