Jurnal Aplikasi Statistika & Komputasi Statistik

Clustering Regencies/Cities Vulnerable to Air Pollution in the Java Island: Fuzzy Geographically Weighted Clustering

Arya Candra Kusuma — 2025-12-31

Introduction/Main Objectives: Air pollution has become a critical global concern with substantial effects on human health and the environment. Background Problems: Java Island in Indonesia, recognized for its high population density and industrial activities, necessitates focused effort in resolving this issue. Novelty: While air pollution research has been enormous, there has been no effort to cluster regencies or cities on Java Island utilizing spatially-based data. This research seeks to cluster regencies and cities on Java Island according to air pollution levels and to compare geodemographic and non-geodemographic clustering methodologies. Research Methods: This study employs secondary data regarding air pollution, obtained from the Openweather API. This study employs a geodemographic clustering technique, namely fuzzy geographically weighted clustering (FGWC), optimized by the artificial bee colony (ABC) algorithm. Finding/Results: The study findings indicate that the geodemographic clustering method ABCFGWC surpasses Fuzzy C-Means (FCM) according to the TSS (Tang-Sun-Sun) index. The data reveal that the Greater Jakarta or Jabodetabek area and its adjacent territories are more susceptible to air pollution. The findings of this study are expected to enhance the spatial planning and mapping of air pollution management strategies on Java Island.

Adding MSNBURR-IIa Distribution to MultiBUGS

Eliana Putri Ramadani — 2025-12-31

Introduction/Main Objectives: The MSNBurr-IIa distribution is a neo-normal distribution designed to fit right-skewed data better. This article aims to integrate the MSNBurr-IIa distribution into MultiBUGS, thereby enabling Bayesian estimation of its parameters. Background Problems: Markov Chain Monte Carlo (MCMC) is a popular method for Bayesian computations, although its implementation is frequently challenging. MultiBUGS, a statistical tool that uses the BUGS language, is used to make this easier. Novelty: This paper details integrating the MSNBurr-IIa distribution into MultiBUGS, allowing for estimating its parameters. The module's effectiveness is demonstrated through its application on both simulated data and regional economic growth data of Indonesian districts/cities in 2021. Research Methods: The MSNBurr-IIa module was developed using five steps: requirement, design, development, testing, and implementation in simulation and real-world data. It was built with Blackbox Component Builder, an integrated development environment (IDE) for the Component Pascal programming language. Finding/Results: The findings confirm that MultiBUGS, with the MSNBurr-IIa module, successfully estimates the distribution’s parameters across various datasets.

Implementation of Twofold HB Beta SAE Model to Estimate Out-of-School Children with Disabilities in Indonesia

Aisha Maharani — 2025-12-31

Introduction/Main Objectives: The high percentage of out-of-school children with disabilities in Indonesia reveals a significant gap in educational participation. Background Problems: Due to the absence of disability-focused surveys, accurate data are only available at the national level, which is insufficient to represent regional conditions. Novelty: With the increasing demand for small area data, this study estimates the percentage of out-of-school children with disabilities at the provincial and district levels simultaneously, using small area estimation (SAE). Research Methods: This study applies SAE using a twofold subarea-level model with a Hierarchical Bayes (HB) beta approach, covering all 34 provinces and 514 districts/cities in Indonesia. This model was developed using data from the National Socio-Economic Survey (Susenas) and the Village Potential Statistics (Podes). Finding/Results: The twofold HB beta SAE model achieves higher precision than direct estimation, as shown by lower relative standard errors (RSE) across regions. Furthermore, spatial patterns indicate that the percentage of out-of-school children with disabilities is mostly between 35.36% and 45.34%, with clusters concentrated in Kalimantan and Papua.

Spatial Heterogeneity of Food Security in Indonesia: Unpacking the Roles of Technology and Democracy Index

Ditto Satrio Wicaksono — 2025-12-31

Introduction/Main Objectives: Food security is a key concern for all countries, especially Indonesia. Technological development and democratic quality are vital for sustainable food security. This study aims to determine the impact of technology and democracy on food security. Background Problems: The relationship between food security and these two factors remains uncertain. Moreover, the extant literature on the spatial impacts on food security yields results that are inconclusive. Novelty: This study offers a comprehensive depiction of the impact of spatial relationships between variables, with a particular focus on the quality of democracy and technology, on the multidimensionality of food security. Research Methods: A spatial lag model is applied to ascertain the impact of technological and democratic on multidimensional food security using data from 34 provinces in 2022. Finding/Results: The results reveal significant spatial dependence in Indonesia’s food security. Technological development and democratic quality positively and significantly affect food security, while urbanization and food crop land expansion show negative and positive effects, respectively. Spatial spillover accounts for approximately 37%–38% of the total impact of each explanatory variable. These findings suggest that technology adoption, democratic strengthening, and interprovincial collaboration are crucial for improving food security.

Forecasting Farmer Exchange Rate (FER) in Southeast Sulawesi Province Using Cheng’s Fuzzy Time Series Method

Rastina — 2025-12-31

Introduction/Main Objectives: This study aims to forecast the Farmer Exchange Rate (FER) in Southeast Sulawesi Province for 2024 as a basis for short-term economic assessment and policy-related analysis. Background Problems: FER is a key indicator of farmers’ purchasing power and agricultural welfare; however, its monthly dynamics are characterized by fluctuations and uncertainty, making conventional forecasting methods less effective in capturing its behavior. Novelty: This study contributes by implementing the the Fuzzy Time Series (FTS) Cheng approach for FER forecasting in Southeast Sulawesi, emphasizing its suitability for handling vagueness and nonlinear patterns inherent in agricultural economic indicators. Research Methods: The analysis utilizes monthly secondary FER data obtained from BPS-Statistics of Southeast Sulawesi Province, covering the period from January 2014 to December 2023. Forecast accuracy is evaluated using the Mean Absolute Percentage Error (MAPE). Finding/Results: The forecasting results indicate that the FER values for January, February, and March 2024 are each estimated at 105.93. The model achieved a MAPE of 0.3027%, corresponding to an accuracy level of 99.6973%, which places the forecasting performance in the “excellent” category.

Satellite Imagery for Classification Analysis of Abrasion Areas on Panaitan, Banten

Ghaffar Ismail — 2025-12-31

Introduction/Main Objectives: Abrasion causes severe environmental degradation and socio- economic losses, Waton and Karang Gundul Islands have already subsided due to erosion, posing risks to Panaitan Island, a national park that also faces deforestation, infrastructure development, and vegetation loss which may intensify abrasion. Background Problems: Limited spatial data on coastal abrasion in Panaitan Island hampers effective monitoring and management, highlighting the need for spatially explicit analysis. Novelty: This study identified and classified abrasion-prone areas on Panaitan Island (a rarely exposed island) with rarely variables which have impactful indices such as MVI, TCI, and LSWI. Research Methods: Landsat 8 and Sentinel-2 imagery from 2018 and 2023 were analyzed to assess changes in vegetation, mangroves, surface temperature, and soil moisture. Random Forest, Support Vector Machine, and Logistic Regression were employed to classify abrasion-prone areas. Finding/Results: The analysis revealed signs of abrasion covering 2.04 km², with Random Forest achieving the highest accuracy (82.23%) and NDVI as the most influential variable; abrasion was mainly associated with declining forest and mangrove cover, soil moisture showed weak correlation, while moderate surface temperature had a positive effect. Preventive measures such as reforestation and mangrove rehabilitation are recommended to mitigate risks and ensure long-term environmental sustainability.

Estimating Economic Activity Using Geospatial Big Data in East Java, Indonesia: Relative Spatial GDP Index Approach

Rifqi Ramadhan — 2025-12-31

Introduction/Main Objectives: GRDP serves as a fundamental indicator for assessing regional economic performance in Indonesia and plays a critical role in development planning. Background Problems: Conventional GRDP measurement in Indonesia relies on survey-based approaches, which are time-consuming, costly, and provide limited spatial detail. Novelty: This study introduces a Relative Spatial GDP Index (RSGI) constructed from geospatial big data such as remote sensing and point of interest (POI) to estimate GRDP more granular in East Java. This approach represents the first geospatial data driven GRDP index developed at such fine spatial resolution in Indonesia. Research Methods: Four weighting schemes were applied to generate RSGI variations, which were then evaluated through regression modeling against official GRDP. They are equal weight, pearson correlation, spearman correlation, and principal component analysis (PCA). Finding/Results: The RSGI PCA produced the best performance (RMSE = 0.73047; MAE = 0.48185; MAPE = 7.00%; R² = 0.7618). PCA weight outperformed other weight by capturing shared variance and generating objective weights that better represent spatial economic intensity. The RSGI PCA demonstrates a strong and significant correlation with GRDP at the sub-district level and provides a robust tool for fine-scale economic estimation.

Integrating Multi-Criteria Decision Analysis and Machine Learning for Fine-Scale Mapping of Safe Drinking Water Access in Bengkulu Province, Indonesia

Andrew Maruli Tua Tampubolon — 2025-12-31

Introduction/Main Objectives: This study aims to develop a 1 km × 1 km level estimation model of safe drinking water access using multisource satellite imagery, point of interest (POI), and aquifer productivity maps. Background Problems: There is a lack of alternative data sources for estimating safe drinking water access that are cost-, time-, and labor-efficient while maintaining high accuracy and frequent updates. Novelty: This study integrates Multi-Criteria Decision Analysis (MCDA) and machine learning methods to estimate and map safe drinking water access at a 1 km × 1 km resolution. Research Methods: Multisource geospatial data were used to construct the model. Within the MCDA approach, the Weighted Product Model (WPM) was employed to develop the Safe Drinking Water Access Index (SDWAI). Meanwhile, the machine learning regression algorithms Adaptive Boosting Regression (ABR) and Gradient Boosting Regression (GBR) were applied to estimate safe drinking water access at a fine spatial scale. The study was conducted in Bengkulu Province, Indonesia. Finding/Results: WPM yielded the best MCDA performance ( = 0.3699, RMSE = 10.6566, MAE = 9.5427, MAPE = 0.1405), while ABR showed the best machine learning performance ( = 0.4361, RMSE = 10.0813, MAE = 8.3750, MAPE = 0.1333).