Bulletin of Applied Mathematics and Mathematics Education

Data assimilation for predicting the dynamics of acute respiratory infections using the ensemble Kalman filter

2026-02-25T03:29:26+00:00

Acute respiratory infection (ARI) is one of the most pressing public health problems due to its high transmission rate and the potential to cause significant pressure on health services. This study applies the Ensemble Kalman Filter (EnKF) method to predict the spread of ARI with a three-compartment population model, namely Susceptible (S), Exposed (E), and Infected (I). This study shows that the EnKF method can predict the spread of ARI well. The number of ensembles used affects the level of accuracy. The EnKF provides accurate predictions of the dynamics of ARI spread, making it relevant as a scientific basis in the formulation of data-based mitigation strategies. It can provide a scientific basis for policymakers to formulate accurate and measurable preventive measures.

Geographically weighted panel regression using Haversine distance for mapping sustainable development goals

2026-05-13T08:01:32+00:00

The Sustainable Development Goals (SDGs) vary widely across Asian countries, indicating that the factors driving SDGs achievement may vary by location. Global models may miss these local variations, so this study used Geographically Weighted Panel Regression (GWR Panel), a method that estimates separate regression coefficients for each geographical location. The GWR Panel in this study was used to capture spatially varying SDGs determinants across 46 Asian countries from 2015 to 2024. This study also compares four Adaptive Kernel functions (Gaussian, Exponential, Bisquare, Tricube) with Haversine distances, as kernel choice directly affects which neighboring countries influence each local coefficient estimate, where applying the incorrect kernel to spatially heterogeneous data can lead to biased local estimates. The best kernel was selected using Cross-Validation (CV). The Adaptive Exponential Kernel produced the lowest CV value (81.686), compared to Adaptive Gaussian (83.128), Bisquare (84.485), and Tricube (85.095), confirming it as the most accurate kernel for this data. The results identified 16 distinct country groups, demonstrating that SDGs determinants vary across Asia. Education, gender, economic growth, infrastructure, environment, institutions, and partnerships are universally important. Meanwhile, water, health, hunger, and climate show the greatest regional variation. SDGs policies should be adjusted to local contexts.

Clustering Indonesian Provinces Based on Welfare Level Using Several Validity Indices

2026-04-23T07:39:05+00:00

One of the national development goals is to increase the level of community welfare. There are several aspects that influence the level of welfare, namely population, health, education, housing, social, employment, consumption, and poverty. This research aims to group provinces in Indonesia based on their level of welfare so that the government can determine appropriate policies in the context of economic recovery and improving the welfare of the Indonesian people. The data used are indicators of provincial welfare levels in Indonesia in 2022 from the Central Statistics Agency. Data is grouped into 3 clusters based on welfare level, namely high (C1), medium (C2), and low (C3) using the K-Means and Fuzzy C-Means methods. Based on the results of the validity test, it is known that ththe best method is the K-Means method with Euclidean distance using the parameter k = 3, the resulting DBI value is 0.989 and the C-Index is 0.076, where this value is better than those of the Fuzzy C-Means method. It is hoped that the results can provide information regarding the characteristics of provinces in Indonesia based on welfare level indicators and become a reference for the government in improving welfare in Indonesia.

Clustering of productivity in the food, plantation, and farm sectors in Mamasa Regency using K-Means clustering analysis

2026-05-11T07:18:48+00:00

The productivity of the food, plantation, and farm sectors is the main driver of the economy in Mamasa Regency. However, there are significant disparities between subdistricts, requiring targeted development strategies. This study aims to group 17 subdistricts based on the productivity characteristics of the three main sectors using the K-Means Clustering method. The secondary data analyzed includes productivity, production, planted area, and the number of farmers/ranchers for each sector. The research stages include descriptive analysis, identification of leading commodities, and classification of subdistricts based on the productivity of each sector. The analysis results divide the subdistricts into three main clusters with different characteristics. Cluster 1 (High Productivity) is dominated by leading subdistricts, such as Pana for cattle farming, Nosu for Arabica coffee, and Tabulahan for patchouli. Cluster 2 (Medium Productivity) includes subdistricts with balanced performance in several commodities, while Cluster 3 (Low Productivity) consists of subdistricts that still face challenges in land optimization, production, and cultivation efficiency. The implication of this study is cluster-based policy recommendations. Local governments are advised to implement specific strategies, such as developing cattle breeding centers in Pana, processing Arabica coffee in Nosu, and patchouli industry in Tabulahan. For low-productivity clusters, interventions are directed at improving infrastructure, access to inputs, and technological assistance. With this evidence-based strategy, local potential can be optimized, regional disparities reduced, and economic growth in Mamasa Regency can be more inclusive and sustainable.

A Mathematical Model of Malaria Transmission Dynamics with Multi-Stage Infection and Dual Treatment Pathways

2026-04-30T08:24:42+00:00

Malaria remained a complex global public health challenge due to the interplay between biological transmission and human treatment-seeking behavior. This study developed a deterministic mathematical model incorporating two levels of infection severity (mild and severe) and dual treatment pathways, namely herbal and medical treatment. The model was formulated as a system of nonlinear ordinary differential equations and analyzed using the Next Generation Matrix method to derive the basic reproduction number , , while local stability was examined using the Routh–Hurwitz criterion. The results showed that the disease-free equilibrium was locally asymptotically stable when , indicating the eventual elimination of the disease. Sensitivity analysis revealed that mosquito mortality and transmission rates were the most influential parameters affecting disease spread. Numerical simulations further demonstrated that increasing early-stage treatment, particularly herbal treatment for mild infections, significantly reduced and limited progression to severe cases. These findings highlighted the critical role of early treatment-seeking behavior combined with effective vector control in reducing malaria transmission and supporting long-term elimination strategies.