Jurnal Statistika dan Aplikasinya

Front Matter Jurnal Statistika dan Aplikasinya Volume 8 Issue 1, June 2024

Journal Editor JSA — 2024-06-30

Back Matter Jurnal Statistika dan Aplikasinya Volume 8 Issue 1, June 2024

Journal Editor JSA — 2024-08-16

Application of Geographically Weighted Regression for Modeling the Poverty Cases in Kalimantan, Indonesia

Noerul Hanin — 2024-06-30

Poverty is one of the most global issues that remains a concern worldwide, including in Indonesia. Indonesia is among the top 100 poorest countries in the world, ranked 73rd, to be exact. Government wants to decrease the national poverty rate, as outlined in the 2021-2024 National Medium-Term Development Plan, with the expected percentage of poor people in Indonesia on 2024 being 6.5 to 7 percent. Unfortunately, the hope for a reduction in the poverty rate has not been achieved in several regions, such as in 4 out of 5 provinces in Kalimantan. Therefore, the analyzing factors causing poverty in the Kalimantan region is conducted using the Geographically Weighted Regression method in order to give clearly information for government to decrease the poor rate in this region. GWR (Geographically Weighted Regression) is an extension of the regression method. The equation parameters for each observation location differ from one location to another. The weighting function used were fixed gaussian, fixed bisquare, fixed tricube, adaptive gaussian, adaptive bisquare, and adaptive tricube. Based on R2 and AIC value, the best model is the model with fixed tricube function. The R2 score is 0.8952, while the AIC score is 155.83. The GWR model is better than OLS or global regression model. Thus, spatial analysis to see the factors affecting the percentage of poor people in each regency and city in Kalimantan, Indonesia has been successfully carried out.

Characteristics of Provinces in Indonesia Based on JKN Indicator Outcomes by Gaussian Mixture Model with Expectation-Maximization Algorithm and Biplot

Dania Siregar — 2024-06-30

Indonesia, an archipelago with a population of 257.77 million in 2022, faces significant challenges in enhancing the quality of life to improve human resource productivity. This study aims to identify provincial characteristics in Indonesia based on the outcomes of the Jaminan Kesehatan Nasional (JKN) program from 2019 to 2021. Using a Gaussian Mixture Model (GMM) with the Expectation Maximization (EM) algorithm, we cluster 34 provinces based on 14 health indicators. The data were obtained from the BPJS website and included variables such as access to health services, program effectiveness, and service quality. Our methodology allows for clustering provinces with similar health outcomes and analyzing the unique indicators for each cluster using biplot analysis.The results indicate significant variation in cluster membership across the years. In 2019, three clusters were identified, with cluster sizes of 16, 12, and 6 provinces. In 2020, the optimum model also had three clusters, but with different member distributions: 24, 7, and 3 provinces. By 2021, four clusters emerged with sizes of 9, 16, 3, and 6 provinces. These findings highlight the dynamic nature of health outcomes across Indonesia's provinces and suggest the need for tailored policy interventions to improve the JKN program's effectiveness.The study's limitations include the reliance on available BPJS data and the assumption that the selected health indicators comprehensively represent the JKN program's impact. This research's novelty lies in its use of advanced clustering techniques to provide a nuanced understanding of regional health disparities in Indonesia, which can inform more targeted and effective health policies.

Sentiment Analysis of Public Opinion on Handling Stunting in Indonesia using Random Forest

Ariska Fitriyana Ningrum — 2024-06-30

The issue of stunting is important to address, as it has the potential to affect the human resource potential and is related to health levels, and even child mortality. The Indonesian government targets to reduce the stunting rate to 14 percent by 2024 through an accelerated stunting reduction program as an effort to improve the nutritional status of the society and also reduce the prevalence of stunting or stunted children. Understanding public sentiment towards the stunting initiative is crucial for policymakers and stakeholders to design effective interventions and allocate resources efficiently. This study aims to analyze public sentiment related to stunting in Indonesia, which impacts children's growth and development. Through the use of sentiment analysis techniques, this study aims to understand public perceptions and attitudes towards the issue of stunting, evaluating whether the general sentiment is positive, negative or neutral. The results of this analysis are expected to provide useful insights for policymakers and health practitioners in designing and implementing more effective strategies to address the issue of stunting. This study conducted sentiment analysis from crawled Twitter data, showing positive and negative sentiments of the public regarding stunting handling in Indonesia. Furthermore, classification analysis using random forest was conducted and resulted in an accuracy score of 97.5%. The model is good enough but, we suggest trying other algorithms in further research.

Panel Data Regression Modelling Factors Affecting Participation in Indonesia's National Health Insurance Contribution Assistance Program

Vera Maya Santi — 2024-06-30

National Health Insurance (JKN) is a health insurance system that provides health protection to individuals who have paid premiums (non-PBI) or whose premiums are paid by the government (PBI). JKN participation in Indonesia from 2018 to 2021 has consistently increased, particularly for JKN PBI, yet the program has not achieved the target of 100% membership coverage globally or nationally. A deeper understanding of the factors influencing JKN is crucial to enhance the program's effectiveness and formulate targeted policies. This study aims to analyze the factors affecting the participation of JKN PBI annually in Indonesia. The data used in this study cover the period from 2018 to 2021 and involve 34 provinces in Indonesia, thus forming a panel data structure. A panel data regression model is employed to identify the factors influencing JKN PBI participation. The predictor variables analyzed include income levels, access to health facilities, employed population, and education levels. The results indicate that the most suitable model for determining the factors influencing participation in the National Health Insurance PBI with a significance level of is the Random Effect Model (REM). The simultaneous and partial tests show that the variables of the employed population and education levels significantly impact JKN PBI participation. Therefore, the REM is considered appropriate and effective in explaining the factors influencing participation in the JKN PBI program. This study provides insights that can help in formulating more targeted policies to enhance the effectiveness of the JKN program in Indonesia.

Mapping Domestic and Foreign Tourists in East Java Using C-Means Clustering

Marita Qori'atunnadyah — 2024-06-30

Tourism is a priority sector identified by the government for its potential to drive economic growth, job creation, community development, and regional progress. Although significant, it still requires a detailed mapping of tourist visit patterns to optimize regional tourism potential. This study uses the C-Means Clustering method to categorize districts and cities in East Java based on the number of domestic tourists and foreign tourists. Data from 2018 to 2022 is used to identify different patterns and groups. The methodology involves clustering the data based on similarities in the number of visitors, which provides insight into regional tourism dynamics. The results revealed three main groups of domestic tourists: high, medium, and low-visitation regions. For foreign tourists, five groups were identified, reflecting variations in the level of tourist visits. These groups help understand the distribution and concentration of tourists in different regions, which is important for targeted promotion strategies and efficient resource allocation. A limitation of this study is that it does not go deeper into external factors affecting tourism, such as the COVID-19 pandemic. The originality of this research lies in the application of the C-Means Clustering method to map domestic tourists and foreign tourists in East Java not simultaneously, thus providing valuable insights for policymakers and industry stakeholders to encourage collaboration and innovation in the tourism sector.

Modeling Gross Enrollment Rate for Higher Education in Central Java Province Using Principal Component Geographically Weighted Regression Approach

Mohd Syafrizal — 2024-06-30

Education in a country is a crucial factor in enhancing human resources. The Gross Enrollment Rate for Higher Education is one of the important indicators used by the government to evaluate the development of the education sector, particularly higher education. Social and cultural diversity, as well as geographical influences, result in varying conditions across different regions, leading to each region having its own unique characteristics, known as spatial heterogeneity. Geographically Weighted Regression (GWR) is a method that can address the problem of spatial heterogeneity. Additionally, a common issue encountered in modeling with many independent variables is multicollinearity, which can lead to high variance in regression parameter estimates and invalid conclusions. Principal Component Analysis (PCA) is a dimensionality reduction method that can address multicollinearity. The aim of this research is to model Gross Enrollment Rate in Higher Education in Central Java using GWR, preceded by handling multicollinearity with PCA. Furthermore, this study aims to determine the factors influencing Gross Enrollment Rate in Higher Education in Central Java. The results, using a fixed Gaussian kernel weighting function, indicate that modeling with PCA and GWR performs better than using the Ordinary Least Square (OLS) method alone, yielding an AIC value of 169.43 and a coefficient determination value of 96.2%.

Determinants of Prehypertension and Hypertension Among Indonesian Productive-Age Population Using Ordinal Logistic Regression With Non-Proportional Odds Model

Muhammad Almas Yafi’ — 2024-06-30

Non-communicable diseases (NCDs) account for the majority of deaths in Indonesia, especially cardiovascular diseases. The main risk factor for cardiovascular disease is hypertension, which is also a disease with the highest increase in prevalence compared to other NCDs. The risk of cardiovascular disease doubles for every increase in blood pressure of 20/10 mmHg and starts from a pressure of 115/75 mmHg. Therefore, a prehypertension category was introduced which also aims to prevent the development of hypertension. The prevalence of hypertension in the majority of the productive age population (15─64 years) has exceeded the RPJMN 2015─2019 target, even though Indonesia is currently in the demographic bonus period. This research aims to obtain a general picture and factors that influence blood pressure status in the productive age population in Indonesia in 2018. The data used comes from Riskesdas which is integrated with Susenas in 2018. The results of the ordinal logistic regression analysis (non-proportional odds model) show that consumption of fruit and vegetables, consumption of fatty foods, consumption of salty foods, consumption of seasonings, alcohol consumption, smoking status, physical activity, education level, employment status, poverty status, residence, age, gender, and body mass index have a significant effect on blood pressure status in the productive age population.

Comparative Analysis of Classical Methods with Machine Learning Algorithm on Survival Classification of Heart Failure Patients

Sa'idah Zahrotul Jannah — 2024-06-30

Cardiovascular disease is a global threat and is the main cause of death worldwide. More than 17.9 million people died from heart and blood vessel problems. Most of these deaths, around 80%, occurred in countries with low or middle economies, including Indonesia. This research aims to find the most accurate and efficient model for classifying cardiovascular disease data so that cardiovascular disease can be detected early.

This research uses heart failure patient data with predictor and response variables. The response variable has two categories such as passed away and alive. Moreover, predictor variables are obtained from the patient’s behavioral risk factors. Data preprocessing was done before the modeling and divided into 0% training and 20% testing data. Modeling in training data was done with multiple algorithms such as Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbor (KNN), and Support Vector Machine (SVM). Each model was evaluated with metrics such as Accuracy, Precision, and Recall obtaining the best model.

This study found that the use of all research variables in the classification analysis leads to a decrease in classification performance, so this study used SelectKBest with a total of 8 significant variables. Furthermore, the Random Forest algorithm with optimal parameters using entropy criterion and a maximum depth of 8 is the method with the most optimal performance, achieving a precision of 90.51% for the 'alive' category, recall of 88.27% for 'alive', the precision of 88.55% for 'deceased', recall of 90.74% for 'deceased', training accuracy of 89.51%, AUC of 0.895, and testing accuracy of 87.80%, placing it in the category of good classification.

Although this research is limited to medical records and behavioral risk factors of heart failure patients to classify patient survival resilience, it addressed data imbalance, employed feature selection, and compared multiple algorithms to provide insights into their effectiveness for this specific classification task and improve model efficiency.

A Multivariate Approach: Forecasting Jakarta Composite Using Prophet Facebook

Arum Handini Primandari — 2024-06-30

The Jakarta Composite Index (JCI, Composite Stock Price Index / IHSG) presents the average share price movement of companies listed on the Indonesia Stock Exchange (BEI/IDX), which can reflect the stock market performance. JCI forecasting can provide benefits for investors regarding risk management. On the other hand, gold is a low-risk asset with no credit risk and maintains its value over time. During the pandemic, gold prices increased significantly while stock prices decreased sharply, so gold prices can be used as a regressor in forecasting the JCI. Researchers obtained historical data on the JCI and gold prices (dollars/ounce) from January 1, 2018, to December 31, 2022. The approach used in this research is multivariate in the Prophet model. The Prophet model uses a procedure to estimate time series data based on an additive model with trends that can be adjusted for annual, weekly, and daily seasonality. Based on the analysis results, the Prophet's multivariate approach is the best method for predicting the JCI compared to the univariate approach. The parameters used in the model are as follows: yearly seasonality, multiplicative seasonality mode, seasonality prior scale, namely 0.5, and changepoint prior scale, namely 0.001. The Mean Absolute Percentage Error (MAPE) obtained from the model is 2.78%.

Factor Analysis of Dengue Hemorrhagic Fever

Nisriina Nazhiifah — 2024-06-30

Dengue Hemorrhagic Fever (DHF) is a disease caused by the dengue virus. DHF cases have always been a serious problem every year in Central Java. This study aimed to determine the factors that cause Dengue Hemorrhagic Fever (DHF) in the province of Central Java because cases of DHF in the region become a serious problem every year. The method used for this research is Principal Component Analysis and Factor Analysis using secondary data from the Central Java Provincial Health Office in 2018. The results of the analysis show that 3 factors are causing DHF, namely the population participation factor in health, sanitation factors, and clean drinking water factors. This shows the importance of environmental education to increase population awareness in terms of healthy living and local government intervention is needed in environmental health projects, namely sanitation and clean drinking water This research only uses seven variables that are considered relevant, other variables that may also have an influence are not included in this analysis. This research focuses on a particular year that shows a decreasing trend in cases. This approach offers a fresh and distinct perspective on understanding the dynamics of dengue fever and the factors that contribute to its reduction.

Clustering Municipality of Dengue Hemorrhagic Fever Typologies in Central Java

Fareka Erdien — 2024-06-30

Cases of dengue hemorrhagic fever (DHF) generally occur in areas with high temperatures. Central Java Province, Indonesia is one of the regions that has high temperatures, making it vulnerable to dengue cases. The study aimed at grouping DHF endemic areas in Central Java needs to be done to assist the government in determining policies to control or prevent DHF. The cluster analysis method used in this study is Average Linkage. The results showed that there were 3 clusters formed. Cluster A is a cluster with the characteristics of having the highest average percentage of households that have access to safe drinking water. Cluster B is the cluster with the highest average number of protected springs. While cluster C is dominant in 4 factors with the highest average, namely the percentage of households that behave in a clean and healthy life, the percentage of healthy homes, the number of Polindes (Village Maternity Hut), and the percentage of households that have access to proper sanitation. Clusters A and B are clean water type and Cluster C is a sanitation type, where clean water and sanitation are both indicators of environmental health. Therefore, environmental health is closely related to the presence of dengue fever in a community environment. The determination of three clusters was based on the chosen method and criteria. Other methods or criteria might suggest a different optimal number of clusters. The findings are specific to Central Java Province and may not be generalizable to other regions with different environmental and social contexts. In this case, it is necessary to pay attention to the community for environmental health in order to overcome or prevent the occurrence of DHF. where clean water and sanitation are both indicators of environmental health. Therefore, environmental health is closely related to the presence of dengue fever in a community environment.