Multi-modal approach for COVID-19 detection using coughs and self-reported symptoms

COVID-19 (Coronavirus Disease of 2019) is one of the most challenging healthcare crises of the twenty-first century. The pandemic causes many negative impacts on all aspects of life and livelihoods. Although recent developments of relevant vaccines, such as Pfizer/BioNTechmRNA, AstraZeneca, or Moderna, the emergence of newvirus mutations and their fast infection rate yet pose significant threats to public health. In this context, early detection of the disease is an important factor to reduce its effect and quickly control the spread of pandemic. Nevertheless, many countries still rely on methods that are either expensive and time-consuming (i.e., Reverse-transcription polymerase chain reaction) or uncomfortable and difficult for self-testing (i.e., Rapid Antigen Test Nasal). Recently, deep learning methods have been proposed as a potential solution for COVID-19 analysis. However, previous works usually focus on a single symptom, which can omit critical information for disease diagnosis. Therefore, in this study, we propose a multi-modal method to detect COVID-19 using cough sounds and self-reported symptoms. The proposed method consists of five neural networks to deal with different input features, including CNN-biLSTM for MFCC features, EfficientNetV2 for Mel spectrogram images, MLP for self-reported symptoms, C-YAMNet for cough detection, and RNNoise for noise-canceling. Experimental results demonstrated that our method outperformed the other state-of-the-art methods with a high AUC, accuracy, and F1-score of 98.6%, 96.9%, and 96.9% on the testing set.

Predictive Analytics of Stock Market as a Time Series

This paper presents a new methodology and a comparative study using past stock market data that can help businesses take investing or divesting decisions in critical situations in the future. These may be like the COVID-19 pandemic, where market volatility is extremely high, thus creating an urgent need for better decision support systems to minimise loss and ensure better profits. The results of the study are based on the comparison of different configurations of ARIMAX, Prophet, LSTM and Bidirectional LSTM Models trained on historical NSE data. By understanding the correlation and variations in the data processing and model training parameters, we have successfully proposed a LSTM neural network model training and optimising method which could successfully help businesses take both long and short term profitable decisions before and after big financial and market crises with a respective accuracy of 98.60 percent and 96.97 percent. © 2022 IEEE.

A Comparison of LSTM and BiLSTM for Forecasting the Air Pollution Index and Meteorological Conditions in Jakarta

Jakarta is the capital city of Indonesia where air pollution becomes one of the problems that must be properly handled. The historical data of the air pollution index is beneficial for developing models for forecasting future values. One of the advantages of forecasting air pollution is to help people to arrange future plans to reduce the dangerous effect on health. Analyzing a record of meteorological conditions can be used to understand climate change. This paper reports the comparison of Long Short Term Memory (LSTM) and Bidirectional LSTM (BiLSTM) models for multivariate forecasting of the air pollution index and meteorological conditions in Jakarta. It also informs the performance of those algorithms for forecasting the observed variables before and during the Coronavirus disease (Covid-19) outbreak to analyze the effect of the pandemic on the environment. The experiments use a historical time series dataset from 2010-2021. The experimental results show that LSTM and BiLSTM work well to forecast PM10, temperature, humidity, and wind speed. In this case study, there are no significant differences in the performance of LSTM and BiLSTM. © 2022 IEEE.

Prediction and assessment of the impact of COVID-19 lockdown on air quality over Kolkata: a deep transfer learning approach.

The present study focuses on the prediction and assessment of the impact of lockdown because of coronavirus pandemic on the air quality during three different phases, viz., normal periods (1 January 2018-23 March 2020), complete lockdown (24 March 2020-31 May 2020), and partial lockdown (1 June 2020-30 September 2020). We identify the most important air pollutants influencing the air quality of Kolkata during three different periods using Random Forest, a tree-based machine learning (ML) algorithm. It is found that the ambient air quality of Kolkata is mainly affected with the aid of particulate matter or PM (PM10 and PM2.5). However, the effect of the lockdown is most prominent on PM2.5 which spreads in the air of Kolkata due to diesel-driven vehicles, domestic and commercial combustion activities, road dust, and open burning. To predict urban PM2.5 and PM10 concentrations 24 h in advance, we use a deep learning (DL) model, namely, stacked-bidirectional long short-term memory (stacked-BDLSTM). The model is trained during the normal periods, and it shows the superiority over some supervised ML models, like support vector machine, K-nearest neighbor classifier, multilayer perceptron, long short-term memory, and statistical time series forecasting model autoregressive integrated moving average. This pre-trained stacked-BDLSTM is applied to predict the concentrations of PM2.5 and PM10 during the pandemic situation of two cases, viz., complete lockdown and partial lockdown using a deep model-based transfer learning (TL) approach (TLS-BDLSTM). Transfer learning aims to utilize the information gained from one problem to improve the predictive performance of a learning model for a different but related problem. Our work helps to demonstrate how TL is useful when there is a scarcity of data during the COVID-19 pandemic regarding the drastic change in concentration of pollutants. The results reveal the best prediction performance of TLS-BDLSTM with a lead time of 24 h as compared to some well-known traditional ML and statistical models and the pre-trained stacked-BDLSTM. The prediction is then validated using the real-time data obtained during the complete lockdown due to COVID second wave (16 May-15 June 2021) with different time steps, e.g., 24 h, 48 h, 72 h, and 96-120 h. TLS-BDLSTM involving transfer learning is seen to outperform the said comparing methods in modeling the long-term temporal dependency of multivariate time series data and boost the forecast efficiency not only in single step, but also in multiple steps. The proposed methodologies are effective, consistent, and can be used by operational organizations to utilize in monitoring and management of air quality.

RSCOEWR: Radical-Based Sentiment Classification of Online Education Website Reviews

Online education is becoming more and more popular with the development of the Internet. In particular, due to the COVID-19 pandemic, many countries around the world are increasing the popularity of online education, which makes the research on sentiment classification of course reviews of online education websites an important research direction in natural language processing tasks. Traditional sentiment classification models are mostly based on English. Unlike English, Chinese characters are based on pictograms. Radicals of Chinese characters can also express certain semantics, and characters with the same radical often have similar meanings. Therefore, RSCOEWR, a word-level and radical-level based sentiment classification model for course reviews of Chinese online education websites is proposed, which solves the problem of data sparsity of reviews by feature extraction of multiple dimensions. In addition, a deep learning model based on CNN, BILSTM, BIGRU and Attention is constructed to solve the problem of high dimension and assigning the same attention to context of traditional sentiment classification model. Extensive comparative experiment results show that RSCOEWR outperforms the state-of-the-art sentiment classification models, and the experimental results on public Chinese sentiment classification datasets prove the generalization ability of RSCOEWR.

A Deep Learning Approach for Recognizing Covid-19 from Chest X-ray using Modified CNN-BiLSTM with M-SVM

At present, Covid-19 is posing serious intimidation to students, doctors, scientists and governments all around the world. It is a single-stranded RNA virus with one of the enormous RNA genomes, and it is constantly changing through mutation. Sometimes this mutation results in a new variant. Research showed that people who come in touch with this virus mostly they are infected with lung illness. So, recognizing Covid-19 from a Chest X-ray is one of the best imaging techniques. But another issue arises when it shows that other diseases like viral pneumonia, lung opacity are also had common symptoms like as Covid-19 and these problems also can be detected from chest X-ray images. So, in this research, we proposed a deep learning approach combining Modified Convolutional Neural Network (M-CNN) and Bidirectional LSTM (BiLSTM) with an Multi-Support Vector Machine (M-SVM) classifier for detecting Covid-19, Viral Pneumonia, Lung-Opacity and normal chest. We used the COVID-19-Radiography-Dataset to assess the results of our proposed system and compared the result with some other existing systems which show our proposed system is better than others. The accuracy of classification using the proposed method is 98.67%. © 2022 IEEE.

Pandemic-Aware Electric Load Forecasting: A Multitask Bidirectional LSTM/CNN Model

COVID-19 pandemic has brought major uncertainty in load forecasting. Enforcing and relaxing lockdown rules, infection numbers, and the changing habits of people are the main causes of this uncertainty. Electric load forecasting maintains the balance between electric supply and demand. It also assists electric utilities in pricing their services, planning, and managing their infrastructure. This paper proposes two pandemic-aware load forecasting models (i) a city-level model, applied on the cities of Ottawa and Toronto, predicting hourly load using weather and pandemic-related features including population mobility and the number of daily COVID-19 infections, and (ii) a second open-source model forecasting quarter-hourly residential-level loads using weather and population mobility features for the city of Pune in India. Both models utilize multitask learning to jointly learn and predict future electric loads. The quarter-hourly model uses Bi-directional Long Short-Term Memory (LSTM) to learn from COVID's specific features, and a Convolutional Neural Network (CNN) to learn from the historical load data before the pandemic. The multitask nature of the model allows for incorporating multiple datasets with different numbers of features. The residential-level multitask model allowed for learning from long-term data before COVID-19 using weather features, short-term load data, and the mobility data. Multitask learning has also enabled the use of two datasets with different numbers of features due to the lack of mobility data pre-COVID. © 2022 IEEE.

Attention-based bidirectional LSTM with embedding technique for classification of COVID-19 articles

The epidemic of COVID-19 has thrown the planet into an awfully tricky situation putting a terrifying end to thousands of lives;the global health infrastructure continues to be in significant danger. Several machine learning techniques and pre-defined models have been demonstrated to accomplish the classification of COVID-19 articles. These delineate strategies to extract information from structured and unstructured data sources which form the article repository for physicians and researchers. Expanding the knowledge of diagnosis and treatment of COVID-19 virus is the key benefit of these researches. A multi-label Deep Learning classification model has been proposed here on the LitCovid dataset which is a collection of research articles on coronavirus. Relevant prior articles are explored to select appropriate network parameters that could promote the achievement of a stable Artificial Neural Network mechanism for COVID-19 virus-related challenges. We have noticed that the proposed classification model achieves accuracy and micro-F1 score of 75.95% and 85.2, respectively. The experimental result also indicates that the propound technique outperforms the surviving methods like BioBERT and Longformer.

Identification of Optimal Model for Multi-Class Classification of COVID Tweets

Covid-19 is the worst-hit pandemic that has affected humankind to date. It sent all major nations around the globe into lockdowns for at least half of 2020. The lockdown started to increase unrest in the population, and even some of them started sharing the emotions infused by the unrest and lockdown over social media platforms in the form of posts, stories, articles. The emotions that underlie those posts can be categorized into three categories positive, negative and neutral, and the individual posts can be classified into respective labels. We considered one of the social platforms' Twitter and collected Twitter tweets. The dataset included the text from the tweet along with emotion. The dataset was pre-processed, including removing stop words from the dataset, stemming and lemmatizing the words from tweets text. Our work focused on various models that can be used to analyze sentiment and classification. The work includes implementing standard classification models like Naive-Bayes multinomial Classifier, K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision Tree Classifier, Logistic Regression, Deep Learning models - Long short-term memory (LSTM), Gated recurrent unit (GRU), Bidirectional long-short term memory (Bidirectional LSTM), Bidirectional Encoder Representations from Transformers (BERT). The results from all these models are compared and tried to establish the most efficient model based on accuracy. The BERT model outperformed all other methods when compared to other models developed using Machine Learning (ML) and Deep Learning (DL) techniques. © 2022 Bharati Vidyapeeth, New Delhi.

Sentiment analysis system for COVID-19 vaccinations using data of Twitter

COVID-19 vaccination topic has been a hot topic of discussions on social media platforms wondering its effectiveness against the SARS-COV-2 virus. Twitter is one of the social media platforms that people widely lunched to express and share their thoughts about different issues touching their daily life. Though many studies have been undertaken for COVID-19 vaccine sentiment analysis, they are still limited and need to be updated constantly. This paper conducts a system for COVID-19 vaccine sentiment analysis based on data extracted from Twitter platform for the time interval from 1st of January till the 3rd of Sep. 2021, and by using deep learning techniques. The introduced system proposes to develop a model architecture based on a deep bidirectional long short-term memory (LSTM) neural network, to analyze tweets data in the form of positive, neutral, and negative. As a result, the overall accuracy of the developed model based on validation data is 74.92%. The obtained outcomes from the sentiment analysis system on collected tweets-data of COVID-19 vaccine revealed that neutral is the prominent sentiment with a rate of 69.5%, and negative sentiment has less rate of tweets reached 20.75% while the positive sentiment has a lesser rate of tweets reached of 9.67%. © 2022 Institute of Advanced Engineering and Science. All rights reserved.

A Deep Learning Approach to Predict COVID-19 Through Cough Analysis Using CNN-BiDirectional LSTM

COVID-19 is deadly contagious and with new variants forming every day, people are at great risk. 29.9 million cases have been reported in India so far and the range goes up to 179 million cases worldwide. Due to the limited number of test centers and the ever-increasing cases, the equipment and the lab technicians are getting outnumbered. They are also at constant risk of getting infected. The scientists have been working on analyzing coughs using machine learning and this technology has been successful in analyzing different types of coughs and classifying them into respective categories. Using this, a deep learning model is created that analyses the recorded cough sample and classifies them into their respective category. A CNN-Bidirectional LSTM is used to create this model and run it on the Covid-sounds dataset provided by Cambridge University. The Covid-sounds dataset has both breathing and cough samples of positive, negative, and other respiratory diseases which might alter or cause cough. This data is pre-processed and used as cough samples for the model. This model has outperformed other models which used the same dataset. © 2022 IEEE.

Adapt-Kcr: a novel deep learning framework for accurate prediction of lysine crotonylation sites based on learning embedding features and attention architecture.

Protein lysine crotonylation (Kcr) is an important type of posttranslational modification that is associated with a wide range of biological processes. The identification of Kcr sites is critical to better understanding their functional mechanisms. However, the existing experimental techniques for detecting Kcr sites are cost-ineffective, to a great need for new computational methods to address this problem. We here describe Adapt-Kcr, an advanced deep learning model that utilizes adaptive embedding and is based on a convolutional neural network together with a bidirectional long short-term memory network and attention architecture. On the independent testing set, Adapt-Kcr outperformed the current state-of-the-art Kcr prediction model, with an improvement of 3.2% in accuracy and 1.9% in the area under the receiver operating characteristic curve. Compared to other Kcr models, Adapt-Kcr additionally had a more robust ability to distinguish between crotonylation and other lysine modifications. Another model (Adapt-ST) was trained to predict phosphorylation sites in SARS-CoV-2, and outperformed the equivalent state-of-the-art phosphorylation site prediction model. These results indicate that self-adaptive embedding features perform better than handcrafted features in capturing discriminative information; when used in attention architecture, this could be an effective way of identifying protein Kcr sites. Together, our Adapt framework (including learning embedding features and attention architecture) has a strong potential for prediction of other protein posttranslational modification sites.

Prediction and Analysis of Covid-19 Positive Cases Using Deep Learning Model

At the end of December 2019, the COVID-19 virus was the initial report case in China Wuhan City. On March 11, 2020. The Department of Health (WHO) announced COVID-19, a global pandemic. The COVID-19 spread rapidly out all over the world within a few weeks. We will propose to develop a forecasting model of COV-19 positive case predict outbreak in Pakistan using Deep Learning (DL) models. We assessed the main features to forecast patterns and indicated The new COVID-19 disease pattern in Pakistan and other countries of the world. This research will use the deep learning model to measure several COVID-19 positive case reports in Pakistan. LSTM cell to process time-series data forecasts is very efficient. Recurrent neural network processes to handle time-dependent and involve hidden layers are confirmed and predict positive cases and weekly cases reported in the future. Bidirectional LSTM (Bi-LSTM) processes data and information in one direction to predict and analyze the weekly 6-9 days readily forecast the number of positive cases of COVID-19 © 2021 IEEE.

A SSA-Based Attention-BiLSTM Model for COVID-19 Prediction

The Corona Virus Disease 2019 (COVID-19) has widely spread over the world and comes up with new challenges to the research community. Accurately predicting the number of new infections is essential for optimizing available resources and slowing the progression of such diseases. Long short-term memory network (LSTM) is a typical method for COVID-19 prediction in deep learning, but it is difficult to extract potentially important features in time series effectively. Thus, we proposed a Bidirectional LSTM (BiLSTM) model based on the attention mechanism (ATT) and used the Sparrow Search Algorithm (SSA) for parameter tuning, to predict the daily new cases of COVID-19. We capture the information in the past and future through the BiLSTM network and apply the attention mechanism to assign different weights to the hidden state of BiLSTM, enhance the ability of the model to learn vital information, and use the SSA to optimize the critical parameters of the model for matching the characteristics of COVID-19 data, enhance the interpretability of the model parameters. This study is based on daily confirmed cases collected from six countries: Egypt, Ireland, Iran, Japan, Russia, and the UK. The experimental results show that our proposed model has the best predictive performance among all the comparison models. © 2021, Springer Nature Switzerland AG.