Forecasting cardiovascular disease mortality using artificial neural networks in Sindh, Pakistan

BMC Public Health, Journal Year: 2025, Volume and Issue: 25(1)

Published: Jan. 4, 2025

Abstract Cardiovascular disease (CVD) is a leading cause of death and disability worldwide, its incidence prevalence are increasing in many countries. Modeling CVD plays crucial role understanding the trend cases, evaluating effectiveness interventions, predicting future trends. This study aims to investigate modeling forecasting mortality, specifically Sindh province Pakistan. The civil hospital Nawabshah area province, Pakistan, provided data set used this study. It time series dataset with actual cardiovascular mortality cases from 1999 2021 included. analyzes forecasts deaths Pakistan using classical models, including Naïve, Holt-Winters, Simple Exponential Smoothing (SES), which have been adopted compared machine learning approach called Artificial Neural Network Auto-Regressive (ANNAR) model. performance both models ANNAR model has evaluated key indicators such as Root Mean Square Deviation Error, Absolute Error (MAE), Percentage (MAPE). After comparing results, it was found that outperformed all selected demonstrating quantifying burden concludes best-selected among competing for province. provides valuable insights into impact interventions aimed at reducing can assist formulating health policies allocating economic resources. By accurately policymakers make informed decisions address public issue effectively.

Language: Английский

---

Forecasting stock prices using a novel filtering-combination technique: Application to the Pakistan stock exchange

AIMS Mathematics, Journal Year: 2024, Volume and Issue: 9(2), P. 3264 - 3288

Published: Jan. 1, 2024

<abstract><p>Traders and investors find predicting stock market values an intriguing subject to study in exchange markets. Accurate projections lead high financial revenues protect from risks. This research proposes a unique filtering-combination approach increase forecast accuracy. The first step is filter the original series of prices into two new series, consisting nonlinear trend long run stochastic component using Hodrick-Prescott filter. Next, all possible filtered combination models are considered get forecasts each with linear time forecasting models. Then, results combined extract final forecasts. proposed technique applied Pakistan's daily price index data January 2, 2013 February 17, 2023. To assess methodology's performance terms model consistency, efficiency accuracy, we analyze different set ratios calculate four mean errors, correlation coefficients directional Last, authors recommend testing for additional complicated future achieve highly accurate, efficient consistent forecasts.</p></abstract>

Language: Английский

---

Evaluating the forecasting performance of ensemble sub-epidemic frameworks and other time series models for the 2022–2023 mpox epidemic

Royal Society Open Science, Journal Year: 2024, Volume and Issue: 11(7)

Published: July 1, 2024

During the 2022-2023 unprecedented mpox epidemic, near real-time short-term forecasts of epidemic's trajectory were essential in intervention implementation and guiding policy. However, as case levels have significantly decreased, evaluating model performance is vital to advancing field epidemic forecasting. Using laboratory-confirmed data from Centers for Disease Control Prevention Our World Data teams, we generated retrospective sequential weekly Brazil, Canada, France, Germany, Spain, United Kingdom, States at global scale using an auto-regressive integrated moving average (ARIMA) model, generalized additive simple linear regression, Facebook's Prophet well sub-epidemic wave n-sub-epidemic modelling frameworks. We assessed forecast mean squared error, absolute weighted interval scores, 95% prediction coverage, skill scores Winkler scores. Overall, framework outcompeted other models across most locations forecasting horizons, with unweighted ensemble performing best frequently. The spatial-wave frameworks considerably improved relative ARIMA (greater than 10%) all metrics. Findings further support epidemics emerging re-emerging infectious diseases.

Language: Английский

---

Modeling and Analysis of Monkeypox Outbreak Using a New Time Series Ensemble Technique

Axioms, Journal Year: 2024, Volume and Issue: 13(8), P. 554 - 554

Published: Aug. 14, 2024

The coronavirus pandemic has raised concerns about the emergence of other viral infections, such as monkeypox, which become a significant hazard to public health. Thus, this work proposes novel time series ensemble technique for analyzing and forecasting spread monkeypox in four highly infected countries with virus. This approach involved processing first cumulative confirmed case address variance stabilization, normalization, stationarity, nonlinear secular trend component. After that, five single models three proposed are used estimate filtered series. accuracy is evaluated using typical mean errors, graphical evaluation, an equal statistical test. Based on results, it found that efficient accurate way forecast cases top world entire world. Using best model, made next 28 days (four weeks), will help understand disease associated risks. information can prevent further enable timely effective treatment. Furthermore, developed be diseases future.

Language: Английский

---

Multi-step ahead ozone level forecasting using a component-based technique: A case study in Lima, Peru

AIMS environmental science, Journal Year: 2024, Volume and Issue: 11(3), P. 401 - 425

Published: Jan. 1, 2024

<abstract><p>The rise in global ozone levels over the last few decades has harmed human health. This problem exists several cities throughout South America due to dangerous of particulate matter air, particularly during winter season, making it a public health issue. Lima, Peru, is one ten with worst air pollution. Thus, efficient and precise modeling forecasting are critical for concentrations Lima. The focus on developing models anticipate concentrations, providing timely information adequate protection environmental management. work used hourly O$ _{3} $ data metropolitan areas multi-step-ahead (one-, two-, three-, seven-day-ahead) forecasts. A multiple linear regression model was represent deterministic portion, four-time series models, autoregressive, nonparametric autoregressive moving average, nonlinear neural network were describe stochastic component. various horizon out-of-sample forecast results considered suggest that proposed component-based technique gives highly consistent, accurate, gain. may be expanded other districts different regions even level assess efficacy approach. Finally, no analysis been undertaken using estimation Lima manner.</p></abstract>

Language: Английский

---

Optimal Features Selection in the High Dimensional Data based on Robust Technique: Application to Different Health Database

Heliyon, Journal Year: 2024, Volume and Issue: 10(17), P. e37241 - e37241

Published: Sept. 1, 2024

Bio-informatics and gene expression analysis face major hurdles when dealing with high-dimensional data, where the number of variables or genes much outweighs samples. These difficulties are exacerbated, particularly in microarray data processing, by redundant that do not significantly contribute to response variable. To address this issue, selection emerges as a feasible method for identifying most important genes, hence reducing generalization error classification algorithms. This paper introduces new hybrid approach combining Signal-to-Noise Ratio (SNR) score robust Mood median test. The test is beneficial impact outliers non-normal skewed since it may successfully identify significant changes across groups. SNR measures significance gene's comparing gap between class means within-class variability. By integrating both these approaches, suggested aims find tasks. objective study evaluate effectiveness combination choosing optimal genes. A P-value consistently identified each using score. dividing value its P-value, Md calculated. Genes high signal-to-noise ratio have been considered favorable due their minimal noise influence importance. verify selected utilizes two dependable techniques: Random Forest K-Nearest Neighbors (KNN). algorithms were chosen track record completing categorization-related performance evaluated metrics: reduction accuracy. metrics offer an in-depth assessment how well improve accuracy consistency. According findings, put out here outperforms conventional methods datasets has lower rates. There considerable improvements specific exposed KNN classifiers. outcomes demonstrate technique might be helpful tool processes bioinformatics.

Language: Английский

---

Investigating Mpox Strain Dynamics Using Computational and Data-Driven Approaches

Viruses, Journal Year: 2025, Volume and Issue: 17(2), P. 154 - 154

Published: Jan. 23, 2025

This study explores Mpox transmission dynamics using a mathematical and data-driven epidemiological model that incorporates two viral strains, Clade I II. The includes pathways between humans mammals divides the human population into susceptible, exposed, infectious, hospitalized, recovered groups. Weekly data from WHO for Spain, Italy, Nigeria, DRC 2022 to 2024 are used validation via non-linear least-squares fitting, with performance assessed by Root Mean Squared Error (RMSE). We conduct time-series analysis detect trends anomalies in cases, scenario simulations examining strain-specific basic reproduction number (R0). fit is compared statistical fits emphasize importance of developing strain. Mathematical confirms model’s key properties, including positivity, boundedness, equilibrium stability. Results underscore varying infection proportions R0. combines rigor empirical provide valuable insights offers framework understanding multi-strain pathogens diverse populations. simulation indicate an increase effective contact rate leads dominance prevalent Clades each country. Based on these findings, we recommend implementation strategies aimed at reducing control spread virus strains.

Language: Английский

---

MRpoxNet: An enhanced deep learning approach for early detection of monkeypox using modified ResNet50

Digital Health, Journal Year: 2025, Volume and Issue: 11

Published: Jan. 1, 2025

To develop an enhanced deep learning model, MRpoxNet, based on a modified ResNet50 architecture for the early detection of monkeypox from digital skin lesion images, ensuring high diagnostic accuracy and clinical reliability. The study utilized Kaggle MSID dataset, initially comprising 1156 augmented to 6116 images across three classes: monkeypox, non-monkeypox, normal skin. MRpoxNet was developed by extending 177 182 layers, incorporating additional convolutional, ReLU, dropout, batch normalization layers. Performance evaluated using metrics such as accuracy, precision, recall, F1 score, sensitivity, specificity. Comparative analyses were conducted against established models like ResNet50, AlexNet, VGG16, GoogleNet. achieved 98.1%, outperforming baseline in all key metrics. demonstrated superior robustness distinguishing lesions other conditions, highlighting its potential reliable application. provides robust efficient solution detection. Its performance suggests readiness integration into workflows, with future enhancements aimed at dataset expansion multimodal adaptability diverse scenarios.

Language: Английский

---

Non-stationary fuzzy time series modeling and forecasting using deep learning with swarm optimization

International Journal of Machine Learning and Cybernetics, Journal Year: 2025, Volume and Issue: unknown

Published: March 4, 2025

Language: Английский

---

A stacked ensemble approach for symptom-based monkeypox diagnosis

Computers in Biology and Medicine, Journal Year: 2025, Volume and Issue: 191, P. 110140 - 110140

Published: April 8, 2025

The recent monkeypox outbreak has raised global health concerns. Caused by a virus, it is characterized symptoms such as skin lesions. Early detection critical for treatment and controlling its spread. This study uses advanced machine learning deep techniques, including Tab Transformer, Long Short-Term Memory, XGBoost, LightGBM, Stacking Classifier, to predict the presence of virus based on patient symptoms. performance these models evaluated using accuracy, precision, recall, F1-score metrics. experiments reveal that Classifier significantly outperforms other models, achieving an accuracy 87.29 %, precision 86.12 recall 87.47 F1 score 87.89 %. Additionally, applying Conditional Tabular GAN generate synthetic data helps address imbalance issues, further improving model robustness. These results highlight proposed approach's potential timely, accurate detection, aiding in effective disease management control.

Language: Английский

---