Enthusiastic : International Journal of Applied Statistics and Data Science

Modeling the Prevalence of Stunting in Indonesia Using Quantile Regression

Farida Hayati — 2025-04-26

Stunting is a condition where a child’s height is under the average height of their age. Stunting will have an impact on the quality of human resources. The 2022 Indonesian Nutrition Status Survey reported that the prevalence of stunting in Indonesia reached 21.6%. This number decreased compared to the previous year. However, it remains below the government’s planned target of 14%. Therefore, appropriate methods are needed to model and identify the factors with the most significant impact on the data for each region studied. This research modeled the stunting problem using quantile regression. Quantile regression has several advantages, including the fact that it can be used on data with an inhomogeneous distribution and is not affected by outliers. The results showed that variables that had a significant effect on the prevalence of stunting using 0.95 quantile regression included babies receiving exclusive breast milk, percentage of family planning participants, percentage of households with access to adequate sanitation, low birth weight (LBW) babies, and percentage of toddlers who have Maternal and Child Health (MCH) books. It is hoped that this research can be utilized to carry out appropriate interventions to reduce the prevalence of stunting that occurs in Indonesia.

Temperature Prediction in Norway Using GRUs: A Machine Learning Approach

Andrie Pasca Hendradewa — 2025-04-26

Accurate temperature forecasting in Norway is significant for environmental stewardship and disaster management, in addition to providing essential support for critical sectors, including agriculture, urban development, and energy resource management. This study employed the gated recurrent unit (GRU) to augment the precision of temporal temperature forecasts. After that, it was used to project temperatures for seven days. The dataset, obtained from https://www.yr.no/nb, comprised records of minimum and maximum temperatures spanning from February 1, 2018, to December 31, 2024. The data was partitioned, with 80% allocated for training and 20% designated for testing. Utilizing a training regimen of 20 epochs alongside a three-day lookback interval, the model attained R² scores of 0.82 for minimum temperature predictions and 0.86 for maximum temperature forecasts. These results underscore the GRU model’s capacity to accurately capture daily temperature variations and produce dependable predictions. Given its commendable performance on training and testing datasets, the GRU model is particularly suitable for temperature forecasting.

Claim Reserving Estimation Using the Double Chain Ladder Method with the Bootstrap Approach

Tiffany Audrey Josepa — 2025-04-26

The claim reserve is the amount of funds the insurance company must set aside to pay claims reported by policyholders. Estimation of claim reserves is carried out as a preventive step for failed payment if the reported claim exceeds the insurance company’s capacity. The estimation of claim reserves in this study was performed using the double chain ladder method with a bootstrap approach. The data used was in the form of a run-off triangle of claim counts and claim amounts presented in incremental and cumulative form. The purpose of this research was to determine the estimated value of reported but not settled (RBNS) and incurred but not reported (IBNR) claim reserves through the bootstrap application on the double chain ladder method. After performing the double chain ladder calculation, the estimated RBNS claim reserves amounted to 6,828,456,000 and the IBNR amounted to 3,714,144,000. Meanwhile, using the bootstrap approach, the RBNS claim reserve estimate was 6,777,539,000 and the IBNR was 3,741,979,000. With the conclusion that the greater the nominal claim reserve allocated, the lower the chance of the company going bankrupt.

Analysis of Multinomial Logistics Regression on the Students Faith Data

Mutijah — 2025-04-26

It is essential for the prospective teacher students of Islamic education to have a high faith level because it will influence their behavior. In addition, it positively impacts their social life. The level of a person’s faith will have a different impact; hence, it needs to be measured. The faith concepts and their measurement have been widely developed recently. One of them is a faith concept, which is built by two dimensions or variables, i.e., belief and behavior or feeling. Both variables can fluctuate between very high, high, moderate, low, or very low, each influencing faith. Students who study in an Islamic Religious Education study program and at the same time attend Islamic boarding school are predicted to have high faith. This paper aims to describe the level of student faith and find a suitable multinomial logistic regression model through analysis of its method. Data was collected using questionnaires filled out by 52 students. The results showed that the percentage of students’ faith levels with very high level was 5.8%, high was 36.5%, moderate was 38.5%, low was 13.5%, and very low was 5.8%. Meanwhile, the model accuracy was 94.2%.

Spatial Analysis of Earthquake Intensity Distribution in Java Using the Interpolation Method (2022–2024)

Laras Niken Dwi Cahyani — 2025-04-26

Java, situated in the Pacific Ring of Fire, is one of the most seismically active regions in the world, with frequent earthquakes posing significant risks to its dense population and critical infrastructure. This study aimed to analyze the spatial distribution and intensity patterns of earthquakes in Java from 2022 to 2024 using data from the Meteorology, Climatology, and Geophysics Agency (Badan Meteorologi, Klimatologi, dan Geofisika, BMKG). Spatial interpolation techniques—inverse distance weighted (IDW), nearest neighbor, and Thiessen polygon—were applied to evaluate their effectiveness in mapping earthquake intensity patterns. The dataset included the earthquake magnitude, location, and occurrence time, with performance evaluated using mean absolute percentage error (MAPE) and mean absolute error (MAE). Results showed that the nearest neighbor method achieved the highest accuracy (MAPE of 12.27%, MAE of 0.37), followed by IDW, while the Thiessen polygon method demonstrated limited suitability for continuous seismic phenomena. These findings underscore the importance of selecting appropriate interpolation methods for seismic risk mapping, providing actionable insights for disaster preparedness and urban planning in Java.

A Zero-Inflated Ordered Probit Approach to Modeling Household Poverty Levels

Nidya Putri Yudhani — 2025-04-26

This research addressed the limitations of the ordered probit (OP) regression model in handling data that contains an excessive number of zero responses. The zero-inflated ordered probit (ZIOP) model was employed to overcome this issue. This model separates the estimation of structural zeros and ordinal outcomes through two distinct components: a binary probit for zero inflation and an OP for ordered categories. Due to the absence of closed-form solutions, parameter estimation was conducted using the maximum likelihood estimation (MLE) method with the Berndt-Hall-Hall-Hausman (BHHH) iterative algorithm. The analysis was based on 4,067 household-level observations from Indonesia’s National Socio-Economic Survey, incorporating indicators of health, education, and standard of living derived from the multidimensional poverty index (MPI) framework. The result of the Vuong test (4.56) confirmed that the ZIOP model significantly outperformed the conventional OP model for zero-inflated ordinal data. Therefore, the ZIOP model is considered more appropriate for analyzing household poverty classifications with a high prevalence of zero observations.

Evaluation of Biclustering Imputation Methods for Glioblastoma Gene Expression Data

Agatha Silalahi — 2025-04-26

Glioblastoma is a highly aggressive primary brain tumor with a low survival rate. One of the main challenges in analyzing glioblastoma gene expression data is the presence of missing values, which can reduce biclustering accuracy and affect biological interpretation. This research compared six imputation methods k-nearest neighbors (KNN), mean imputation, singular value decomposition, nonnegative matrix factorization, soft impute, and autoencoderon the GSE4290 gene expression dataset with missing values ranging from 5% to 50%. An evaluation using root mean square error (RMSE), mean absolute error (MAE), and structural similarity index measure (SSIM) showed that soft impute provided the best performance at all levels of missing values, with RMSE of 0.0076, MAE of 0.0073, and perfect SSIM of 1.0000 at 50% missing values. Meanwhile, deep learning-based autoencoder experienced significant performance degradation at high missing values. These findings indicate that more complex models are not always superior, and regularization-based approaches like soft impute are more effective in preserving the biological structure of the data. The results of this research contribute to the optimization of imputation strategies to improve the accuracy of biclustering analysis in glioblastoma studies.

Analysis of Industrial Waste Quality Control Using Generalized Variance and Hotelling’s T2 Control Diagram Methods

Isna Hamidah — 2025-04-26

Environmental pollution is an unsettling problem for everyone and the ecosystem which can be caused by poorly managed waste originated from the final output of industrial production processed. It can negatively impact the surrounding environment if it is not handled properly. Therefore, the waste must be processed until it meets the predetermined characteristic standards before being disposed of. Among the actions that can be taken is carrying quality control. This study aims to evaluate and characterize the quality of the waste produced. The methods used were the generalized variances and Hotelling’s T² control charts. The data used for this research was the characteristics of liquid waste from a sugar factory industry, taken from May to September 2023. The quality control results, which were obtained using the generalized Variance control chart, could be statistically controlled after eight improvements. Then, Hotelling’s T² control chart was successfully controlled after one test. The capability index value obtained was > 1, indicating that the quality control process in liquid waste at the Pesantren Baru sugar factory is capable or controlled.