Revealing the nature of soil liquefaction using machine learning

Earth Science Informatics, Journal Year: 2025, Volume and Issue: 18(2)

Published: Jan. 21, 2025

Language: Английский

---

Management of patients with multiple myeloma and COVID-19 in the post pandemic era: a consensus paper from the European Myeloma Network (EMN)

Leukemia, Journal Year: 2023, Volume and Issue: 37(6), P. 1175 - 1185

Published: May 4, 2023

Language: Английский

---

Genetic justification of COVID‐19 patient outcomes using DERGA, a novel data ensemble refinement greedy algorithm

Journal of Cellular and Molecular Medicine, Journal Year: 2024, Volume and Issue: 28(4)

Published: Feb. 1, 2024

Abstract Complement inhibition has shown promise in various disorders, including COVID‐19. A prediction tool complement genetic variants is vital. This study aims to identify crucial complement‐related and determine an optimal pattern for accurate disease outcome prediction. Genetic data from 204 COVID‐19 patients hospitalized between April 2020 2021 at three referral centres were analysed using artificial intelligence‐based algorithm predict (ICU vs. non‐ICU admission). recently introduced alpha‐index identified the 30 most predictive variants. DERGA algorithm, which employs multiple classification algorithms, determined of these key variants, resulting 97% accuracy predicting outcome. Individual variations ranged 40 161 per patient, with 977 total detected. demonstrates utility ranking a substantial number approach enables implementation well‐established algorithms that effectively relevance outcomes high accuracy.

Language: Английский

---

Unraveling Links between Chronic Inflammation and Long COVID: Workshop Report

The Journal of Immunology, Journal Year: 2024, Volume and Issue: 212(4), P. 505 - 512

Published: Feb. 5, 2024

As COVID-19 continues, an increasing number of patients develop long COVID symptoms varying in severity that last for weeks, months, or longer. Symptoms commonly include lingering loss smell and taste, hearing loss, extreme fatigue, "brain fog." Still, persistent cardiovascular respiratory problems, muscle weakness, neurologic issues have also been documented. A major problem is the lack clear guidelines diagnosing COVID. Although some studies suggest due to prolonged inflammation after SARS-CoV-2 infection, underlying mechanisms remain unclear. The broad range COVID-19's bodily effects responses initial viral infection are poorly understood. This workshop brought together multidisciplinary experts showcase discuss latest research on chronic might be associated with sequelae following infection.

Language: Английский

---

Prognosis of COVID-19 severity using DERGA, a novel machine learning algorithm

European Journal of Internal Medicine, Journal Year: 2024, Volume and Issue: 125, P. 67 - 73

Published: March 8, 2024

Language: Английский

---

Artificial intelligence for diagnosis of mild–moderate COVID-19 using haematological markers

Annals of Medicine, Journal Year: 2023, Volume and Issue: 55(1)

Published: July 12, 2023

Objective The persistent spread of SARS-CoV-2 makes diagnosis challenging because COVID-19 symptoms are hard to differentiate from those other respiratory illnesses. reverse transcription-polymerase chain reaction test is the current golden standard for diagnosing various diseases, including COVID-19. However, this diagnostic method prone erroneous and false negative results (10% -15%). Therefore, finding an alternative technique validate RT-PCR paramount. Artificial intelligence (AI) machine learning (ML) applications extensively used in medical research. Hence, study focused on developing a decision support system using AI diagnose mild-moderate similar diseases demographic clinical markers. Severe cases were not considered since fatality rates have dropped considerably after introducing vaccines.Methods A custom stacked ensemble model consisting heterogeneous algorithms has been utilized prediction. Four deep also tested compared, such as one-dimensional convolutional neural networks, long short-term memory networks Residual Multi-Layer Perceptron. Five explainers, namely, Shapley Additive Values, Eli5, QLattice, Anchor Local Interpretable Model-agnostic Explanations, interpret predictions made by classifiers.Results After Pearson's correlation particle swarm optimization feature selection, final stack obtained maximum accuracy 89%. most important markers which useful Eosinophil, Albumin, T. Bilirubin, ALP, ALT, AST, HbA1c TWBC.Conclusion promising suggest

Language: Английский

---

Revealing the nature of cardiovascular disease using DERGA, a novel data ensemble refinement greedy algorithm

International Journal of Cardiology, Journal Year: 2024, Volume and Issue: 412, P. 132339 - 132339

Published: July 3, 2024

Language: Английский

---

Comparing machine learning algorithms to predict COVID‑19 mortality using a dataset including chest computed tomography severity score data

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: July 13, 2023

Since the beginning of COVID-19 pandemic, new and non-invasive digital technologies such as artificial intelligence (AI) had been introduced for mortality prediction patients. The prognostic performances machine learning (ML)-based models predicting clinical outcomes patients mainly evaluated using demographics, risk factors, manifestations, laboratory results. There is a lack information about role imaging manifestations in combination with predictors. purpose present study to develop an efficient ML model based on more comprehensive dataset including chest CT severity score (CT-SS). Fifty-five primary features six main classes were retrospectively reviewed 6854 suspected cases. independence test Chi-square was used determine most important relevant predictors train algorithms. predictive developed eight algorithms J48 decision tree (J48), support vector (SVM), multi-layer perceptron (MLP), k-nearest neighbourhood (k-NN), Naïve Bayes (NB), logistic regression (LR), random forest (RF), eXtreme gradient boosting (XGBoost). accuracy, precision, sensitivity, specificity, area under ROC curve (AUC) metrics. After applying exclusion criteria, total 815 positive RT-PCR final sample size, where 54.85% male mean age population 57.22 ± 16.76 years. RF algorithm accuracy 97.2%, sensitivity 100%, precision 94.8%, specificity 94.5%, F1-score 97.3%, AUC 99.9% best performance. Other ranging from 81.2 93.9% also good mortality. Results showed that timely accurate stratification could be performed ML-based fed by routine data. proposed CT-SS efficiently predict This lead promptly targeting high-risk admission, optimal use hospital resources, increased probability survival

Language: Английский

---

Dual-stage explainable ensemble learning model for diabetes diagnosis

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 126899 - 126899

Published: Feb. 1, 2025

---

An Improved Model for Medical Forum Question Classification Based on CNN and BiLSTM

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(15), P. 8623 - 8623

Published: July 26, 2023

Question Classification (QC) is the fundamental task for Answering Systems (QASs) implementation, and a vital task, as it helps in identifying question category. It plays big role predicting answer to while building QAS. However, classifying medical questions still challenging due complexity of terms. Many researchers have proposed different techniques solve these problems, but some problems remain partially solved or unsolved. With help deep learning technology, various text-processing become much easier solve. In this paper, an improved learning-based model Medical Forum (MFQC) classify questions. model, feature representation performed using Word2Vec, which word embedding model. Additionally, features are extracted from layer based on Convolutional Neural Networks (CNNs). Finally, Bidirectional Long Short Term Memory (BiLSTM) network used features. The BiLSTM analyzes target information then outputs category via SoftMax layer. Our achieves state-of-the-art performance by effectively capturing semantic syntactic input We evaluate CNN-BiLSTM two benchmark datasets compare its with existing methods, demonstrating superiority accurately categorizing forum

Language: Английский

---