Transmission dynamics informed neural network with application to COVID-19 infections

Computers in Biology and Medicine, Journal Year: 2023, Volume and Issue: 165, P. 107431 - 107431

Published: Sept. 1, 2023

Language: Английский

---

Enhancing COVID-19 forecasting precision through the integration of compartmental models, machine learning and variants

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Aug. 19, 2024

Predicting epidemic evolution is essential for making informed decisions and guiding the implementation of necessary countermeasures. Computational models are vital tools that provide insights into illness progression enable early detection, proactive intervention, targeted preventive measures. This paper introduces Sybil, a framework integrates machine learning variant-aware compartmental models, leveraging fusion data-centric analytic methodologies. To validate evaluate Sybil's forecasts, we employed COVID-19 data from several European U.S. states. The dataset included number new recovered cases, fatalities, variant presence over time. We forecasting precision Sybil in periods which there change trend pandemic or appears. Results demonstrate outperforms conventional approaches, being able to forecast accurately changes trend, magnitude these changes, future prevalence variants.

Language: Английский

---

A novel featurization methodology using JaGen algorithm for time series forecasting with deep learning techniques

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 235, P. 121279 - 121279

Published: Aug. 25, 2023

Language: Английский

---

Analyzing distributed Spark MLlib regression algorithms for accuracy, execution efficiency and scalability using best subset selection approach

Multimedia Tools and Applications, Journal Year: 2023, Volume and Issue: 83(15), P. 44047 - 44066

Published: Oct. 17, 2023

Language: Английский

---

Classification and Identification of Infectious COVID-19 Virus Using Deep Learning and Machine Learning Techniques: A Comprehensive Analysis

SN Computer Science, Journal Year: 2024, Volume and Issue: 5(1)

Published: Jan. 8, 2024

Language: Английский

---

A new method based on generative adversarial networks for multivariate time series prediction

Expert Systems, Journal Year: 2024, Volume and Issue: 41(12)

Published: Sept. 5, 2024

Abstract Multivariate time series have more complex and high‐dimensional characteristics, which makes it difficult to analyze predict the data accurately. In this paper, a new multivariate prediction method is proposed. This generative adversarial networks (GAN) based on Fourier transform bi‐directional gated recurrent unit (Bi‐GRU). First, utilized extend features, helps GAN better learn distributional features of original data. Second, in order guide model fully distribution data, Bi‐GRU introduced as generator GAN. To solve problems mode collapse gradient vanishing that exist GAN, Wasserstein distance used loss function Finally, proposed for air quality, stock price RMB exchange rate. The experimental results show can effectively trend compared with other nine baseline models. It significantly improves accuracy flexibility forecasting provides ideas methods accurate industrial, financial environmental fields.

Language: Английский

---

WDCIP: spatio-temporal AI-driven disease control intelligent platform for combating COVID-19 pandemic

Geo-spatial Information Science, Journal Year: 2023, Volume and Issue: unknown, P. 1 - 25

Published: July 4, 2023

The outbreak and subsequent recurring waves of COVID −19 pose threats on the emergency management people's daily life, while large-scale spatio-temporal epidemiological data have sure come in handy epidemic surveillance. Nonetheless, some challenges remain to be addressed terms multi-source heterogeneous fusion, deep mining, comprehensive applications. Spatio-Temporal Artificial Intelligence (STAI) technology, which focuses integrating spatial related time-series data, artificial intelligence models, digital tools provide intelligent computing platforms applications, opens up new opportunities for scientific control. To this end, we leverage STAI long-term experience location-based services work. Specifically, devise develop a STAI-driven infrastructure, namely, WAYZ Disease Control Intelligent Platform (WDCIP), consists systematic framework building pipelines from automatic collection, processing AI-based analysis inference implementation providing appropriate applications serving various scenarios. According platform logic, our work can performed summarized three aspects: (1) integrated system; (2) hybrid GNN-based approach hierarchical risk assessment (as core algorithm WDCIP); (3) social containment. This makes pivotal contribution facilitating aggregation full utilization multiple sources, where real-time human mobility generated by high-precision mobile positioning plays vital role sensing spread epidemic. So far, WDCIP has accumulated more than 200 million users who been served life convenience decision-making during pandemic.

Language: Английский

---

Decision trees for early prediction of inadequate immune response to coronavirus infections: a pilot study on COVID-19

Frontiers in Medicine, Journal Year: 2023, Volume and Issue: 10

Published: Aug. 2, 2023

Introduction Few artificial intelligence models exist to predict severe forms of COVID-19. Most rely on post-infection laboratory data, hindering early treatment for high-risk individuals. Methods This study developed a machine learning model inherent risk symptoms after contracting SARS-CoV-2. Using Decision Tree trained 153 Alpha variant patients, demographic, clinical and immunogenetic markers were considered. Model performance was assessed Delta datasets. Key factors included age, gender, absence KIR2DS2 gene (alone or with HLA-C C1 group alleles), presence 14-bp polymorphism in HLA-G gene, KIR2DS5 KIR telomeric region A/A. Results The achieved 83.01% accuracy 78.57% variant, True Positive Rates 80.82 77.78%, Negative 85.00% 79.17%, respectively. showed high sensitivity identifying individuals at risk. Discussion present demonstrates the potential AI algorithms, combined epidemiologic, COVID-19 facilitating treatment. Further studies are required routine integration.

Language: Английский

---

Short-Term Solar Insolation Forecasting in Isolated Hybrid Power Systems Using Neural Networks

Inventions, Journal Year: 2023, Volume and Issue: 8(5), P. 106 - 106

Published: Aug. 23, 2023

Solar energy is an unlimited and sustainable source that holds great importance during the global shift towards environmentally friendly production. However, integrating solar power into electrical grids challenging due to significant fluctuations in its generation. This research aims develop a model for predicting radiation levels using hybrid system Gorno-Badakhshan Autonomous Oblast of Tajikistan. study determined optimal hyperparameters multilayer perceptron neural network enhance accuracy forecasting. These included number neurons, learning algorithm, rate, activation functions. Since there are numerous combinations hyperparameters, training process needed be repeated multiple times. Therefore, control algorithm was proposed identify stagnation or emergence erroneous correlations training. The results reveal different seasons require hyperparameter values, emphasizing need meticulous tuning machine models creation varying conditions. absolute percentage error achieved mean one-hour-ahead forecasting ranges from 0.6% 1.7%, indicating high compared current state-of-the-art practices this field. one-day-ahead between 2.6% 7.2%.

Language: Английский

---

Forecasting COVID-19 New Cases Using Transformer Deep Learning Model

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Nov. 3, 2023

Abstract Making accurate forecasting of COVID-19 cases is essential for healthcare systems, with more than 650 million as 4 January, 1 making it one the worst in history. The goal this research to improve precision case predictions Russia, India, and Brazil, a transformer-based model was developed. Several researchers have implemented combination CNNs LSTMs, Long Short-Term Memory (LSTMs), Convolutional Neural Networks (CNNs) calculate total number cases. In study, an effort made correctness models by incorporating recent advancements attention-based time-series forecasting. resulting found perform better other existing showed improved accuracy Using data from different countries adapting will enhance its ability support worldwide combat pandemic giving precise projections

Language: Английский

---