Claim Missing Document
Check
Articles

Found 5 Documents
Search
Journal : Jurnal Varian

Pemodelan Indeks Pembangunan Kesehatan Masyarakat Kabupaten/Kota di Pulau Kalimantan Menggunakan Pendekatan Regresi Probit M. Fathurahman
Jurnal Varian Vol 2 No 2 (2019)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/varian.v2i2.382

Abstract

The Public Health Development Index (PHDI) is a composite indicator that describes the progress of health development and is useful for ranking provinces and districts/cities in achieving successful community health development. In addition, the PHDIM can also be used to determine regional priorities that require assistance in improving health development. Based on the publication of the PHDI ranking by the Health Research and Development Agency of the Ministry of Health in 2013, the PHDI ranking for 55 districts/cities in Kalimantan Island varied greatly. So it needs to be studied, examined the factors that influence it. The purpose of this study was to examine the modeling of the factors that influence the PHDI of districts/cities in Kalimantan Island in 2013 using the probit regression approach. The results of this study indicate that the factors that significantly influence the PHDI of districts/cities in Kalimantan Island in 2013 are the human development index and the labor force participation rate.
Clustreing of Province in Indonesia Based on Education Indicators Using K-Medoids Annisa Zuhri Apridayanti; M Fathurahman; Surya Prangga
Jurnal Varian Vol 7 No 2 (2024)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/varian.v7i2.3205

Abstract

Data mining is searching for interesting patterns or information by selecting data using specific techniques or methods. One method that can be used in data mining is K-Medoids. K-Medoids is a method used to group objects into a cluster. This research aimed to obtain the optimal number of clusters using the K-Medoids method based on Davies-Bouldin Index (DBI) validity on education indicators data by province in Indonesia in 2021. The results showed that the optimal number of clusters using the K-Medoids method based on DBI validity is 5 clusters. Cluster 1 consists of 1 province with a higher average dropout rate, average length of schooling, and well-owned classrooms compared to other clusters. Cluster 2 consists of 15 provinces with an average proportion of school libraries lower than Clusters 3 and 4 and higher than Clusters 1 and 5. Cluster 3 consists of 9 provinces with an average proportion of school libraries, proportions of school laboratories, net enrollment rates, and higher school enrollment rates than other clusters. Cluster 4 consists of 8 provinces with a higher average enrollment rate than the other clusters. Cluster 5 consists of 1 province with a higher average repetition rate and student-per-teacher ratio than other clusters.
Evaluating Different K Values in K-Fold Cross Validation for Binary Logistic Regression to Classify Poverty Sinaga, Julia Oriana; Fathurahman, M.; Wahyuningsih, Sri; Hayati, Memi Nor
Jurnal Varian Vol. 8 No. 2 (2025)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/varian.v8i2.4403

Abstract

Data mining is essential for decision-makers to analyze and extract insights from data efficiently. Classification is one of the data mining techniques used to organize data based on its features, helping to identify patterns and make predictions. This study evaluates Binary Logistic Regression (BLR), a type of generalized linear model that suitable for binary outcomes, for classifying poverty depth across Indonesian regencies/cities in 2022, with a focus on the impact of different K values in K-Fold Cross Validation. The dataset includes 514 regencies/cities, with the Poverty Depth Index as the target variable, categorized into high (1) and low (0) levels, using 11 predictor variables. K-Fold Cross Validation was performed with K values of 3, 5, and 10, using accuracy and Area Under Curve (AUC) as evaluation metrics. The mean accuracy values for BLR are 75.7% for K=3, 74.3% for K=5, and 75.1% for K=10. Results show that K=3 offers the highest accuracy in classifying poverty depth in Indonesia, with the lowest standard deviation of 0.03. However, K=10 demonstrates superior discriminative ability in BLR, reflected by a higher AUC value. This study highlights the significant influence of K values in K-Fold Cross Validation on BLR performance.
Mengeksplorasi Masalah Kejahatan dari POV Statistik dengan Regresi Binomial Negatif Dani, Andrea Tri Rian; Fathurahman, M.; Ni'matuzzahroh, Ludia; Putri Permata, Regita; Putra, Fachrian Bimantoro
Jurnal Varian Vol. 8 No. 2 (2025)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/varian.v8i2.4445

Abstract

Criminality is a complex issue in Indonesia that is very important to the government, law enforcement agencies, and society. The underlying causes of Indonesia's crime problem are complex and impacted by various circumstances. The aim of this research is to model the crime problem in Indonesia and determine the influencing factors.  The method used in this research is Negative Binomial Regression. The results of the study show that the negative binomial regression model can be used to model criminal problems because the variance value is more significant than the average. Based on the parameter significance test results, both simultaneously and partially, the open unemployment rate, Gini ratio, average years of schooling, and prevalence of inadequate food consumption significantly affect the crime rate, with an Akaike’s Information Criterion Corrected (AICc) value of 698,098. These findings suggest that addressing economic inequality, unemployment, education, and food security could help reduce crime in Indonesia. Policies aimed at improving job opportunities, reducing income disparity, and enhancing education and food security are crucial in mitigating crime. This study provides valuable insights for policymakers and law enforcement agencies, offering a foundation for more targeted and effective crime prevention strategies. Future research could employ the robust Poisson Inverse Gaussian Regression method to avoid the overdispersion problem. 
An Implementation of K-Medoids Method in Provincial ClusteringBased on Education Indicator Data Apridayanti, Annisa Zuhri; Fathurahman, M; Prangga, Surya
Jurnal Varian Vol. 7 No. 2 (2024)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/varian.v7i2.3205

Abstract

Data mining is searching for interesting patterns or information by selecting data using specific techniques or methods. One method that can be used in data mining is K-Medoids. K-Medoids is a method used to group objects into a cluster. This research aimed to obtain the optimal number of clusters using the K-Medoids method based on Davies-Bouldin Index (DBI) validity on education indicators data by province in Indonesia in 2021. The results showed that the optimal number of clusters using the K-Medoids method based on DBI validity is 5 clusters. Cluster 1 consists of 1 province with a higher average dropout rate, average length of schooling, and well-owned classrooms compared to other clusters. Cluster 2 consists of 15 provinces with an average proportion of school libraries lower than Clusters 3 and 4 and higher than Clusters 1 and 5. Cluster 3 consists of 9 provinces with an average proportion of school libraries, proportions of school laboratories, net enrollment rates, and higher school enrollment rates than other clusters. Cluster 4 consists of 8 provinces with a higher average enrollment rate than the other clusters. Cluster 5 consists of 1 province with a higher average repetition rate and student-per-teacher ratio than other clusters.