Mathematical Journal of Modelling and Forecasting

Application of the K-Means Clustering Algorithm to the Case of Stunting Risk Families in Districts/Cities of West Sumatra Province in 2023

Widiyanti, Fadhilah Fitri — Thu, 19 Dec 2024 00:00:00 +0000

Stunting is one of the indicators of chronic nutritional status that has a long-term effect on child growth; the main contributing factors are households that do not have access to clean drinking water, proper sanitation facilities, and other factors. The adverse effects experienced by stunted children are reduced cognitive ability, learning ability, decreased endurance, and can lead to new diseases such as diabetes, heart disease, and many other diseases. This study uses the K-Means Cluster method to group the Regency / City of West Sumatra Province in 2023 regarding cases of stunting risk families. K-Means Cluster analysis is an analysis used to group data based on similar features or characteristics. From the results of the study, it can be concluded that the clustering of 19 regencies/cities in West Sumatra Province resulted in 2 groups (clusters): cluster 1 consists of 12 regency/city members, and cluster 2 consists of 7 regency/city members. The characteristic results obtained from each cluster formed are cluster 2 shows families with better conditions than cluster 1.

Application of the Generalized Autoregressive Conditional Heteroskedasticity (GARCH) Model in Forecasting the Volatility of Optimal Portfolio Stock Returns of the MNC36 Index

Siska Deswita, Devni Prima Sari — Thu, 19 Dec 2024 00:00:00 +0000

Investment is a capital investment made by investors through the purchase of several stocks that are usually long-term with the hope that investors will benefit from increased stock prices. The most commonly used risk indicator in investing is volatility. Therefore, it is necessary to carry out modeling that can overcome the effects of heteroscedasticity to predict future volatility. Efforts are made to overcome the effects of heteroscedasticity by applying the Generalized Autoregressive Conditional Heterossexicity (GARCH) Model in Forecasting the Volatility of Optimal Portfolio Stock Returns on the MNC36 Index. This type of research is applied research that begins with reviewing the problem, analyzing relevant theories, and reviewing the problem and its application. Based on the results of data analysis using the residual normality test through the Jarque-Bera test, it was obtained that the GARCH model has a normal residual and is not heteroscedasticity so that it can be used as a forecasting model. BNGA shares obtained the most stable forecast results with almost constant volatility, indicating that this stock has the lowest risk compared to BBCA and BMRI stocks.

Modelling the Number of Stunting Cases in Indonesia in 2022 Using Negative Binomial Regression to Address Overdispersion

Cinta Rizki Oktarina, Reza Pahlepi — Thu, 19 Dec 2024 00:00:00 +0000

This study models the incidence of stunting in toddlers in Indonesia in 2022 using negative binomial regression to address the overdispersion issue often present in count data. The Poisson regression model, typically used for count data, showed less accurate results due to the variance exceeding the mean, indicating overdispersion. By adopting a negative binomial regression approach, this study accommodates higher variability in the data, leading to more accurate estimates. The results reveal that the percentage of pneumonia cases and low birth weight are significant factors in stunting incidence. In contrast, other variables, such as complete basic immunization and poverty levels, are insignificant in the final model. The final negative binomial model yielded a lower AIC value than the initial model, indicating an improved model fit, with an R-squared (Nagelkerke's R²) of 50.50%. This study offers enhanced insights into the factors influencing stunting, supporting more targeted health policy decisions to reduce stunting rates in Indonesia.

Application of Principal Component Analysis in Identifying Factors Affecting the Human Development Index

Muhammad Faisal, Fadhilah Fitri, Zilrahmi — Thu, 19 Dec 2024 00:00:00 +0000

This study examines the Human Development Index (HDI) in West Sumatra Province in 2023. The HDI is an essential indicator for measuring the success of efforts to improve the quality of human life. This research aims to identify the key factors that influence the HDI. The HDI is constructed from three fundamental dimensions that indicate human quality of life: health, education, and economy. The factors within each dimension tend to be strongly correlated, as they mutually influence one another, potentially leading to multicollinearity issues. Therefore, an analysis is conducted to reduce the number of original variables into new orthogonal variables while preserving the total variance of the original variables using Principal Component Analysis (PCA). Based on this background, the study applies PCA to address multicollinearity and to identify new, more representative variables. The study findings indicate that the factors influencing the HDI are the education and economic and health welfare indexes.

Alternative Strategies to Eradicate Corruption in Indonesia with Numerical Simulation of 4th Order Runge Kutta Method on Mathematical Models

Deddy Rahmadi, Pipit Pratiwi Rahayu — Thu, 19 Dec 2024 00:00:00 +0000

Corruption remains a critical issue hindering Indonesia's development across various sectors, necessitating innovative approaches to combat it. This study explores alternative strategies to eradicate corruption by leveraging mathematical modeling and numerical simulations. A dynamic system representing corruption propagation is formulated, considering key variables such as enforcement intensity, public awareness, and policy effectiveness. The 4th-order Runge Kutta method simulates the model and analyzes the impact of various strategic interventions over time. The results show that the difference in initial conditions significantly affects the level of corruption, which increases or decreases in a specific time. These findings provide valuable insights for policymakers and stakeholders in designing effective, data-driven anti-corruption strategies, emphasizing the integration of rigorous mathematical tools with socio-political frameworks. The study highlights the potential of numerical simulations as a complementary approach to traditional qualitative analyses in addressing complex societal challenges like corruption.

Variance and Semi-Variance with a Multi-Objective Approach Using the Spiral Optimization Method

Femilya Sri Zulfa, Dina Agustina — Thu, 19 Dec 2024 00:00:00 +0000

In minimizing the risk faced by investors while maximizing returns, it is necessary to study different risk measures for portfolio optimization, namely mean-variance, and semi-variance, so that it can provide a deeper understanding of how each approach works in various market conditions. The mean-variance approach measures risk based on the total variance of portfolio returns. While the semi-variance approach only focuses on downside risk, which is the risk of loss that is more relevant to investors who tend to be conservative. By comparing these two risk measures, investors can understand the trade-offs in choosing a portfolio management strategy. To conduct a study on portfolio optimization, the author uses a multi-objective optimization approach on the mean-variance and semi-variance models, which will be solved with a spiral model. The results of this study are that the spiral model with a simple case that does not involve high dimensions can be solved quickly. However, for high dimensions with significant maximum spread points and iterations, the algorithm in this Matlab programming runs slowly, so it is ineffective in computation. This spiral method is suspected of having several solutions trapped in local minima, or the results obtained have not converged, so the resulting Pareto front is not optimal.