https://repository.auw.edu.bd/handle/123456789/63 Wed, 05 Aug 2026 16:24:29 GMT 2026-08-05T16:24:29Z https://repository.auw.edu.bd/handle/123456789/3379 Deep learning with LSTM networks for dengue disease forecasting and control Nowsher, Tahmid; Arafat, Shaiful Islam; Kamrujjaman, Md. Accurate forecasting of infectious disease dynamics is essential for effective public health planning and early outbreak response. Traditional statistical models and modern deep learning approaches have both been widely used for epidemiological time–series forecasting. Dengue fever remains a major public health concern in Bangladesh, where seasonal outbreaks often produce sharp and unpredictable infection spikes. Reliable forecasting models are therefore important for anticipating infection trends and supporting early intervention strategies. This study investigates the performance of Long Short-Term Memory (LSTM) networks for predicting daily dengue infection counts in Bangladesh using surveillance data from 2010 to 2022. Multiple LSTM architectures were developed by varying the number of LSTM units, lookback window size, and data preprocessing strategies. A fixed time based data split was used, with training data up to 30 September 2021 and testing on subsequent observations to avoid information leakage. The proposed LSTM models were compared with several baseline forecasting approaches, including Naive, Seasonal-Naive, and ARIMA models. Model performance was evaluated using standard forecasting metrics including Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Symmetric Mean Absolute Percentage Error (SMAPE). Experimental results indicate that classical statistical models remain strong benchmarks for dengue forecasting. Among all evaluated models, the ARIMA(2,1,2) model achieved the best predictive performance on the test dataset with MAE of 8.40 and RMSE of 14.72. The best-performing LSTM configuration produced competitive results with MAE of 8.92 and RMSE of 15.87, outperforming simple baseline approaches such as the Naive and Seasonal-Naive models. These findings suggest that while deep learning models are capable of capturing complex temporal patterns, their advantage over well-tuned statistical models is not guaranteed for this dataset. The results highlight the importance of benchmarking deep learning models against classical time series approaches when forecasting epidemiological data. While LSTM networks demonstrate promising performance in modeling nonlinear temporal dynamics, the ARIMA model provided the most accurate predictions for the studied dengue dataset. Fri, 17 Jul 2026 00:00:00 GMT https://repository.auw.edu.bd/handle/123456789/3379 2026-07-17T00:00:00Z https://repository.auw.edu.bd/handle/123456789/3376 Assessing the Trade-Off between Voluntary and Forced Interventions to Control the Emergence of Recurring Pandemics—An Evolutionary Game-Theoretic Modeling Kamrujjaman, Md. In this study, we aim to examine the dynamics of diseases by employing both voluntary and forced control strategies backed by evolutionary game theory (EGT). The impact of quarantine is investigated through our suggested frame- work provided that a partial adoption of voluntary vaccination is observed at the earlier stage. The combined and individual effect of dual preventive pro- visions are visualized through SEIR-type epidemic model. Additionally, the effect of coercive control policies’ efficacy on individual vaccination decision is illustrated through the lens of EGT. We also consider the cost associated with vaccination and quarantine. The numerical simulations shown in our work emphasize how important it is to put quarantine rules in place to stop the spread of infection. These restrictions imposed by the government can be relieving, especially during times when a sizable section of the populace is re- luctant to get vaccinated because of its ineffectiveness or excessive cost. We also show when and under what circumstances one policy works better than the other. How these policies’ compliance rates should be calculated is there- fore becomes a focal point of discussion. We support this claim by producing phase diagrams for three different evolutionary outcomes throughout our in- vestigation and changing the two crucially important pick-up rate parameters, one connected with the quarantine policy and the other is related to the isola- tion policy, in various directions. We then additionally examine the efficacy and cost associated with different policy adaption. This model effectively high- lights the importance of dual provisional safety in understanding public health issues by using the mean-field approximation technique, which aligns with the well-known imitation protocol known as individual-based risk assessment dynamics. Wed, 26 Mar 2025 00:00:00 GMT https://repository.auw.edu.bd/handle/123456789/3376 2025-03-26T00:00:00Z https://repository.auw.edu.bd/handle/123456789/3375 Mathematical Analysis of a Resource-Based Dispersal Model With Gompertz Growth and Optimal Harvesting Kamrujjaman, Md. Gompertz dynamics offer significant applications for the growth of invasive species, cancer modeling, optimal harvesting policies, sustainable yield, and maintaining population levels due to its pattern formation in low-density cases. This paper examines a widely applicable nonhomogeneous diffusive Gompertz law with zero Neumann boundary conditions, where all coefficients are smooth periodic functions. The analytical approach explains the ubiquitous stability of a time-periodic solution and seeks the optimal strategy for harvesting under the Gompertz growth law, potentially generalizing the results for many small organisms, including plants and wild populations. The proposed model successfully investigates the dynamics with and without diffusion. Moreover, the spatio-temporal equation more precisely describes the population’s evolutionary processes using a generalized classical reaction-diffusion equation. Finally, we observe several potential applications, outlining the optimal strategies for real-world scenarios and related fields where optimal harvesting is utilized. Wed, 01 Jan 2025 00:00:00 GMT https://repository.auw.edu.bd/handle/123456789/3375 2025-01-01T00:00:00Z https://repository.auw.edu.bd/handle/123456789/3374 Solutions of Nonlinear Parabolic PDEs: A Novel Technique Based on Galerkin-Finite Difference Residual Corrections Kamrujjaman, Md. Numerical solutions for second-order parabolic partial differential equations (PDEs), specifically the nonlinear heat equation, are investigated with a focus on analyzing residual corrections. Initially, the Galerkin weighted residual method is employed to rigorously formulate the heat equation and derive numerical solutions using third-degree Bernstein polynomials as basis functions. Subsequently, a proposed residual correction scheme is applied, utilizing the finite difference method to solve the error equations while adhering to the associated error boundary and initial conditions. Enhanced approximations are achieved by incorporating the computed error values derived from the error equations into the original weighted residual results. The stability and convergence of the residual correction scheme are also analyzed. Numerical results and absolute errors are compared against exact solutions and published literature for various time and space step sizes, demonstrating the effectiveness and precision of the proposed scheme in achieving high accuracy. Sat, 25 Oct 2025 00:00:00 GMT https://repository.auw.edu.bd/handle/123456789/3374 2025-10-25T00:00:00Z