Steiner et al 2022a - Revision history

Scipediacontent at 10:33, 15 December 2022

2022-12-15T10:33:27Z

Scipediacontent at 10:31, 15 December 2022

2022-12-15T10:31:18Z

Rimni at 12:44, 7 September 2022

2022-09-07T12:44:31Z

Show changes

Rimni at 14:33, 6 September 2022

2022-09-06T14:33:04Z

Rimni at 13:45, 6 September 2022

2022-09-06T13:45:30Z

Rimni: /* 4. Results and discussion */

2022-09-06T13:44:36Z

‎4. Results and discussion

Rimni: /* 4. Results and discussion */

2022-09-06T13:43:43Z

‎4. Results and discussion

Show changes

Rimni at 13:37, 6 September 2022

2022-09-06T13:37:26Z

Show changes

Rimni at 13:32, 6 September 2022

2022-09-06T13:32:08Z

Show changes

Rimni: /* 4.1. Systematic literature review */

2022-09-06T13:31:03Z

‎4.1. Systematic literature review

@@ Line 768: / Line 768: @@
 {| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; width: 70%;"
 |-
-|style="padding:10px;"| [[File:Review_936792395077_5667_Figure 2.svg|500px]]
+|style="padding:10px;"| [[File:Steiner_et_al_2022a_2680_Figure_2.png|500px]]
 |- style="text-align: center; font-size: 75%;"
 | colspan="1" style="padding:10px;"| '''Figure 2'''. Top 54% of journals in the selected papers
@@ Line 779: / Line 779: @@
 {| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; width: 60%;"
 |-
-|style="padding:10px;"| [[File:Review_936792395077_1718_Figure 3.svg|500px]]
+|style="padding:10px;"| [[File:Steiner_et_al_2022a_9166_Figure_3.png|500px]]
 |- style="text-align: center; font-size: 75%;"
 | colspan="1" style="padding:10px;"| '''Figure 3'''. Top 61% of publications with the most used methods
@@ Line 792: / Line 792: @@
 {| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; width: 60%;"
 |-
-|style="padding:10px;"| [[File:Review_936792395077_2204_Figure 4.svg|500px]]
+|style="padding:10px;"| [[File:Steiner_et_al_2022a_3203_Figure_4.png|500px]]
 |- style="text-align: center; font-size: 75%;"
 | colspan="1" style="padding:10px;"| '''Figure 4'''. Top 70% of authors' nationality
 The first three positions, in relation to nationality, refer to the two countries with the highest number of cases of Covid-19 (USA and India) and the country where the virus was identified (China).

@@ Line 95: / Line 95: @@
 {| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; width: auto;max-width: auto;"
 |-
-|style="padding:10px;"|  [[File:Review_936792395077_4578_Figure 1.svg|600px]]
+|style="padding:10px;"|  [[File:Steiner_et_al_2022a_6611_Figure_1.png|600px]]
 |- style="text-align: center; font-size: 75%;"
 | colspan="1" style="padding:10px;"| '''Figure 1'''. Systematic literature review flowchart
 |}
 Based on the initial search criteria, 1,912 articles were identified in the three scientific databases. The most frequently cited articles were then selected, resulting in a total of 130 articles, of which 10 were duplicates and three did not fit the initial search profile (the first lay outside the predefined time interval, the second was an article from the field of pharmaceutics, and the third was not an original article, but a report). These 13 articles were removed in the exclusion of duplicates and search refinement stages, leaving a total of 117 original articles whose contents were analyzed.

@@ Line 801: / Line 801: @@
 The first three positions, in relation to nationality, refer to the two countries with the highest number of cases of Covid-19 (USA and India) and the country where the virus was identified (China).
-=5. Concluding Remarks=
+==5. Concluding Remarks==
-The aim of the present study was to conduct a systematic and bibliometric review of the articles involving the protocol shown in '''Fig. 2''' which, being the most frequently cited in the literature, set out to answer, among other questions, why certain patients are severely affected by the virus while others are asymptomatic. M. Li ''et al.'' [16], for example, commented that age, sex, temperature, humidity, social distancing, smoking, investments in health, level of urbanization and race can influence the severity of the disease. On the other hand, Di Castelnuovo ''et al.'' [97] concluded that impaired renal function, elevated C-reactive protein and advanced age were major predictors of in-hospital death. However, there are many patients who fit these specific conditions, but who present different degrees of aggravation of the disease.
+The aim of the present study was to conduct a systematic and bibliometric review of the articles involving the protocol shown in [[#img-2|Figure 2]] which, being the most frequently cited in the literature, set out to answer, among other questions, why certain patients are severely affected by the virus while others are asymptomatic. M. Li et al. [16], for example, commented that age, sex, temperature, humidity, social distancing, smoking, investments in health, level of urbanization and race can influence the severity of the disease. On the other hand, Di Castelnuovo et al. [97] concluded that impaired renal function, elevated C-reactive protein and advanced age were major predictors of in-hospital death. However, there are many patients who fit these specific conditions, but who present different degrees of aggravation of the disease.
 We firmly believe that the answer to this question is directly found in very recent studies that reveal a possible genetic predisposition to serious cases of Covid-19 [133,134,135]. Researchers discovered that more severe cases of the disease are associated with the low performance of molecules that identify the virus and are inherited from parents: class I human leukocyte antigen (HLA-I) molecules represent the group of molecules responsible for identifying and distinguishing everything that is in the body and what is not. Six HLA-I molecules, found on the surface of all cells, form a unique set for each individual, which is determined by the genes received from the parents [135]. In other words, it is very important to investigate whether there is a direct link between the seriousness of the disease and the performance of HLA-I in the identification of Sars-CoV-2.
 Therefore, it would be very interesting and promising to analyze the genotype of patients who have suffered Covid-19 and healthy people, employing ML techniques and classifying patients, for instance, as serious, moderate, or mild cases.
+<!--==Declaration of competing interest==
 ==Declaration of competing interest==
 None.
 ==Funding sources==
@@ Line 819: / Line 819: @@
 ==References==
-[1] WHO. Coronavirus disease (COVID-19) pandemic. Retrieved from [https://www.who.int/emergencies/diseases/novel-coronavirus-2019 https://www.who.int/emergencies/diseases/novel-coronavirus-2019] 2021.
+[1] WHO. Coronavirus disease (COVID-19) pandemic. Retrieved from [https://www.who.int/emergencies/diseases/novel-coronavirus-2019 https://www.who.int/emergencies/diseases/novel-coronavirus-2019], 2021.
-[2] Piret J., Boivin G. Pandemics Throughout History. Front. Microbiol., 11(January), 2021.
+[2] Piret J., Boivin G. Pandemics throughout history. Front. Microbiol., 15 January, 2021.
-[3] Huremović D. Brief History of Pandemics (Pandemics Throughout History). In Psychiatry of Pandemics, Springer International Publishing, Cham, pp. 7–35, 2019.
+[3] Huremović D. Brief history of pandemics (Pandemics Throughout History). Psychiatry of Pandemics, Springer International Publishing, Cham, pp. 7–35, 2019.
 [4] Carvalho T., Krammer F., Iwasaki A. The first 12 months of COVID-19: a timeline of immunological insights. Nat. Rev. Immunol., 21(4):245–256, 2021.
-[5] WHO. WHO Coronavirus (COVID-19) Dashboard. Retrieved from [https://covid19.who.int/ https://covid19.who.int/] 2021.
+[5] WHO. WHO Coronavirus (COVID-19) Dashboard. Retrieved from [https://covid19.who.int/ https://covid19.who.int/], 2021.
-[6] Shahid O., Nasajpour M., Pouriyeh S., et al. Machine learning research towards combating COVID-19: Virus detection, spread prevention, and medical assistance. J. Biomed. Inform., 117(October 2020):103751, 2021.
+[6] Shahid O., Nasajpour M., Pouriyeh S., et al. Machine learning research towards combating COVID-19: Virus detection, spread prevention, and medical assistance. J. Biomed. Inform., 117:103751, 2021.
 [7] Prabha B., Kaur S., Singh J., Nandankar P., Kumar Jain S., Pallathadka H. Intelligent predictions of Covid disease based on lung CT images using machine learning strategy. Mater. Today Proc., xxxx), 2021.

← Older revision		Revision as of 13:45, 6 September 2022
Line 558:		Line 558:

	Finally, the last 38 of the 117 selected articles that address general subjects involving ML techniques and Covid-19 are listed in [[#tab-4\|Table 4]]. These include, for example, that of Di Castelnuovo et al. [97], who attempted to list those that aimed to identify the characteristics predisposing Covid-19 patients to in-hospital death. For this purpose, the authors used the data of 3,894 patients from 30 clinical centers distributed throughout Italy, who were hospitalized from February 19th to May 23rd, 2020. The authors used the RF technique to achieve their goal. They concluded that impaired renal function, elevated C-reactive protein, and advanced age were major predictors of in-hospital death.		Finally, the last 38 of the 117 selected articles that address general subjects involving ML techniques and Covid-19 are listed in [[#tab-4\|Table 4]]. These include, for example, that of Di Castelnuovo et al. [97], who attempted to list those that aimed to identify the characteristics predisposing Covid-19 patients to in-hospital death. For this purpose, the authors used the data of 3,894 patients from 30 clinical centers distributed throughout Italy, who were hospitalized from February 19th to May 23rd, 2020. The authors used the RF technique to achieve their goal. They concluded that impaired renal function, elevated C-reactive protein, and advanced age were major predictors of in-hospital death.
		+

	<div class="center" style="font-size: 75%;">'''Table 4'''. Works (38) related to Covid-19 that used ML techniques involving general subjects</div>		<div class="center" style="font-size: 75%;">'''Table 4'''. Works (38) related to Covid-19 that used ML techniques involving general subjects</div>

@@ Line 234: / Line 234: @@
 |  style="text-align: left;"|Youan Hospital, Beijing.
 |}
 <div class="justified" style="font-size: 75%;">'''Acronyms (alphabetical): '''AI (Artificial Intelligence); AF (Atrial Fibrillation); AL (Active Learning); ANFIS (adaptive neuro-fuzzy inference system); API (Application Programming Interface); AR (Auto-Regressive Process); ARIMA (Auto-Regressive Integrated Moving Average); BTM (Biterm Topic Model); CDCP (Center for Disease Control and Prevention); CART (Classification And Regression Trees); CXR (Chest X-Ray); CMC (composite Monte-Carlo); CMM (Chinese Materia Medica); CNN (Convolutional Neural Networks); CT (computed tomography); ConvLSTM (Convolutional Long Short-Term Memory); CTree (Conditional Inference Tree); CUBIST (Cubist Regression); DCNN (Deep Convolution Neural Networks); DE (Differential Evolution); DL (deep learning); DT (Decision Trees); EA (Ensemble Algorithm); EBM (Ensemble Bagged Model) Trees); ELM (Extreme Learning Machine); FrMEMs (Fractional Multichannel Exponent Moments); FRI (Fuzzy Rule Induction); GA (Genetic Algorithm); GAN (Generative Adversarial Network); GBA (Gradient Boosting Algorithm); GPR(Gaussian Process Regression); GBM (Gradient Boosted Tree Models); GIWD (Generalized Inverse Weibull distribution); GHOST (Globally Harmonized Observational Surface Treatment); HCA (Hierarchical Clustering Algorithm); IoT (Internet of Things); k-NN (k-Nearest Neighbor); K-ELM (Kernel Extreme Learning Machine); LASSO (least absolute shrinkage and selection operator); LDA (Latent Dirichlet Allocation); LSTM (Long Short-Term Memory); LR (Linear Regression); LoR (Logistic Regression); LOS (Length of Stay); LSTM (Long /Short Term Memory); LR (Linear Regression); ML (Machine Learning); MLDSP (Machine Learning with Digital Signal Processing); MLP (Multilayer perceptron); MLP-ICA (MLP-imperialist competitive algorithm); MODE (Multi-objective Differential Evolution); MNB (Multinomial Naïve Bayes); NLP (Natural Language Processing); NCBI (National Center for Biotechnology Information); NN (Neural Network); PAC (Passive Aggressive Classifier); PCR (Principal Components Regression); PDR-NML (Partial Derivative Regression and Nonlinear Machine Learning); PLS-DA (Partial Least Squares Discriminant Analysis); PLSR (Partial Least Squares Regression); PNN+cf (Polynomial Neural Network with Corrective Feedback); PR (Polynomial Regression); RF (Random Forest); RFR (Random Forest Regression); RIDGE (Ridge Regression); RT (Regression Tree); SARIMA (Seasonal Auto-Regressive Integrated Moving Average); SEIRD (Susceptible, Exposed, Infected, Recovered, and Dead); SIR (Susceptible(P-Infected-Recovered epidemiological model); SLR (Simple Linear Regression); SMOM (Social Mimic Optimization Method); SNA (social network analysis); SUFMACS (SUperpixel based Fuzzy Memetic Advanced Cuckoo Search); SVM (Support Vector Machine); SVR (Support Vector Regression); STN (Spatial Transformer Network); SVC (Support Vector Classifier);  SVR (Support Vector Regression); TClustVID (Clustered Based Proposed Classification and Topics modeling Approach); TCM (traditional Chinese medicine); TWC (Topological Weighted Centroid); USDA ERS (United States Department of Agriculture, Economic Research Service); VGG (Visual Geometry Group based Neural Network); WHO (World Health Organization); VAR (Vector Autoregression); WSIDEA (Weighted Stochastic Data Envelopment Analysis); WSCC (Web of Science Core Collection); XGBoost (Extreme Gradient Boosting).</div>
@@ Line 241: / Line 243: @@
 Gulati et al. [41], for example, presented a comparative analysis of seven ML classifiers, such as Linear Support Vector Classifier (SVC), Perceptron, Passive Aggressive Classifier (PAC), and Logistic Regression (LoR). They used more than 72,000 tweet datasets related to Covid-19 pandemic and achieved an accuracy score higher than 98%. Haupt et al. [42] used interdisciplinary approaches to big data, ML, content analysis, and social network analysis (SNA) to characterize the communicative behavior, conversation themes, and network structures of “Liberate protest” supporters and non-supporters. For this purpose, the authors used unsupervised ML techniques and social network analysis. Praveen et al. [43] conducted their study to analyze Indian citizens’ perceptions of what causes stress, anxiety, and trauma during Covid-19. For this purpose, the authors used ML techniques, more specifically, Natural Language Processing (NLP) in 840,000 tweets. Of the 117 articles analyzed, 13 were in this line of research and are listed in [[#tab-2|Table 2]].
 <div class="center" style="font-size: 75%;">'''Table 2'''. Works (13) related to Covid-19 that used ML techniques and social network tools</div>
@@ Line 316: / Line 319: @@
 Acronyms (alphabetical): See [[#tab-1|Table 1]].
 In turn, 44 of the 117 selected articles involved ML methods to predict a wide range of aspects, such as the number of patients who will be infected or intubated, the trends of the pandemic, the production of a real-time Covid-19 SEIRD (Susceptible, Exposed, Infected, Recovered, and Dead) model, and student performance.
 Amar et al. [54], for example, attempted to investigate the disease to eliminate its effects and, to this end, the authors examined a real database from Egypt, from February 15, 2020, to June 15, 2020. They predicted the number of patients that would be infected and estimated the final size of the pandemic. For this purpose, they applied several regression analysis models. Burdick et al. [55] attempted to predict patients’ need for ventilation to determine a better allocation of resources and prevent emergency intubations and their associated risks. The authors analyzed 197 patients, from five USA health systems between March 24 and May 4, 2020. The patients were enrolled in the REspirAtory Decompensation for the triage of the disease: a prospective studY (READY) clinical trial. Of the 117 articles analyzed, 44, including the two already mentioned above, were in this line of research and are listed in [[#tab-3|Table 3]].
 <div class="center" style="font-size: 75%;">'''Table 3'''. Works (44) related to Covid-19 that directly used ML techniques for prediction</div>

@@ Line 123: / Line 123: @@
 |-style="text-align:center"
 ! style="background-color: #d1e0df;"| Authors (year) !!  style="background-color:#d1e0df;text-align:left;"| Focus of the paper !!  style="background-color: #d1e0df;text-align:left;"| ML Techniques !! style="background-color: #d1e0df;text-align:left;"| Databases
-|-
+|-style="background-color:#ebf5f4;"
 |  style="text-align: center;"|Anastasopoulos et al. (2020) [22]
 |  style="text-align: left;"|Implement an automated software to solve the substantial increase in chest CT admissions