cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
-
Contact Email
-
Phone
-
Journal Mail Official
-
Editorial Address
-
Location
Kota semarang,
Jawa tengah
INDONESIA
Media Statistika
Published by Universitas Diponegoro
ISSN : -     EISSN : 24770647     DOI : -
Core Subject : Science,
Decision Sciences, Operations Research & Management
Arjuna Subject : -
Articles 11 Documents
 1   2 
Search results for , issue "Vol 18, No 1 (2025): Media Statistika" : 11 Documents clear
EVALUATING RANDOM FOREST AND XGBOOST FOR BANK CUSTOMER CHURN PREDICTION ON IMBALANCED DATA USING SMOTE AND SMOTE-ENN Andespa, Reyuli; Sadik, Kusman; Suhaeni, Cici; Soleh, Agus M
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.25-36

Abstract

The banking industry faces significant challenges in retaining customers, as churn can critically affect both revenue and reputation. This study introduces a robust churn prediction framework by comparing the performance of XGBoost and Random Forest algorithms under imbalanced data conditions. The novelty of this research lies in integrating the SMOTE and SMOTE-ENN techniques with machine learning algorithms to enhance model performance and reliability on highly imbalanced datasets. Unlike conventional approaches that rely solely on oversampling or undersampling, this study demonstrates that the hybrid combination of XGBoost and SMOTE provides superior predictive accuracy, stability, and efficiency. Hyperparameter optimization using GridSearchCV was conducted to identify the most effective parameter configurations for both algorithms. Model performance was evaluated using the F1-Score and Area Under the Curve (AUC). The results indicate that XGBoost with SMOTE achieved the best performance, with an F1-Score of 0.8730 and an AUC of 0.9828, showing an optimal balance between precision and recall. Feature importance analysis identified Months_Inactive_12_mon, Total_Trans_Amt, and Total_Relationship_Count as the most influential predictors. Overall, this approach outperforms traditional resampling and modeling techniques, providing practical insights for data-driven customer retention strategies in the banking industry.
COMPARISON OF RANDOM FOREST AND SUPPORT VECTOR MACHINE CLASSIFICATION METHODS FOR PREDICTING THE ACCURACY LEVEL OF MADRASAH DATA Syarip, Dodi Irawan; Notodiputro, Khairil Anwar; Sartono, Bagus
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.37-48

Abstract

This study aims to identify the most effective classification method for predicting the accuracy level of madrasah data with class imbalance. Two machine learning approaches were employed: Random Forest (RF) and Support Vector Machine (SVM). Based on the AUC values, it was concluded that the RF model had a slightly better performance in predicting the accuracy level of the madrasah data, with an average AUC of 62.82, compared to the SVM model, which had an average AUC of 62.33. Among all models, the highest and consistent performance was achieved by the RF model using ROSE techniques. The results of measuring variable importance showed that the predictor variables with the greatest influence in predicting the accuracy level of the madrasah data are the number of students and the student-to-teacher and staff ratio. This finding suggests that school principals and madrasah administrative staff should prioritize ensuring the completeness of student, teacher, and staff data to improve the overall reliability of madrasah data.
RANDOM EFFECTS META-REGRESSION USING WEIGHTED LEAST SQUARES (CASE STUDY: EFFECTIVENESS OF ACCEPTANCE AND COMMITMENT THERAPY IN REDUCING DEPRESSION) Arumningtyas, Felinda; Otok, Bambang Widjanarko; Purnami, Santi Wulan
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.49-60

Abstract

Meta-analysis is a statistical method for synthesizing quantitative data from multiple related studies, yet heterogeneity among studies often complicates interpretation. Meta-regression extends this approach by incorporating study-level covariates to explain variations in outcomes. With the global increase in depression, Acceptance and Commitment Therapy(ACT) has attracted attention as an effective psychological intervention. Therefore, a deeper understanding of the factors that influence its effectiveness across studies is needed. However, to date, only a few meta-analyses have quantitatively examined moderator variables that influence ACT outcomes using a random effects meta-regression approach. This study aims to fill this gap. This study estimated the model parameters using the Weighted Least Squares (WLS) method. Thirty-three published studies testing the effectiveness of ACT in reducing depression were collected from PubMed, Google Scholar, and Science Direct. The homogeneity test results showed significant heterogeneity, supporting the use of a random effects model. The combined effect size of -0.321 indicates that ACT significantly reduces depression levels compared to the control group. Meta-regression analysis revealed that variation in effect size was significantly influenced by differences in the average age of patients and duration of therapy. These findings provide new insights into the conditions and characteristics that make ACT therapy more effective.
COMPARATIVE EVALUATION OF ARIMA AND GRU MODELS IN PREDICTING RUPIAH DOLLAR EXCHANGE RATE Fitrianti, Dwi; Ulfia, Ratu Risha; Notodiputro, Khairil Anwar; Angraini, Yenni; Mualifah, Laily Nissa Atul
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.1-12

Abstract

This study evaluates the effectiveness of the ARIMA (Autoregressive Integrated Moving Average) and GRU (Gated Recurrent Unit) models in forecasting the USD–Rupiah exchange rate. Exchange rate fluctuations influence overall economic stability, making accurate forecasting crucial. Monthly data from January 2001 to March 2024 were analyzed. The ARIMA model, a traditional statistical approach, combines autoregressive (AR), differencing (I), and moving average (MA) components to capture linear patterns, while the GRU model, a deep learning approach, captures nonlinear and complex temporal relationships using update and reset gate mechanisms to retain long-term information. Model performance was evaluated using the Mean Absolute Percentage Error (MAPE). The GRU model achieved a MAPE of 1.74%, lower than the ARIMA model’s 1.94%, and generated a forecast of Rp. 16,399.91 for April 2024, closer to the actual value of Rp. 16,249.00 compared to ARIMA’s Rp. 15,857.68. The findings demonstrate the GRU model’s superior forecasting performance and provide empirical evidence of its effectiveness in modeling volatile exchange rate data, particularly the Rupiah–USD rate.
DETERMINATION OF INSURANCE PREMIUMS FOR CHILI PLANTATION USING THE BLACK-SCHOLES MODEL WITH CLAYTON COPULA APPROACH Sutisna, Sarah; Sukono, Sukono; Napitupulu, Herlina
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.13-24

Abstract

Agriculture is a vulnerable sector to the risk of crop damage due to climate change and other environmental factors. One source of risk in agriculture is rainfall, which significantly affects productivity and farmers’ income. Traditional insurance premium calculations often rely on assumptions of normal distribution and linear dependency, which may not accurately capture the complex and non-linear relationships between climatic and agricultural variables. This research presents a novel contribution to agricultural risk management by applying the Clayton Copula to model the dependency structure between rainfall and chili crop production output in the context of crop insurance pricing. The estimation of Copula parameters was conducted using Maximum Likelihood Estimation, yielding a parameter θ value of -0.1252, which indicates the dependency structure between the variables. The predictive accuracy of the Copula Clayton model was evaluated using the Mean Absolute Error, with a result of 0.01291, demonstrating strong relevance in describing the dependency between precipitation and yield. Furthermore, the research integrates the Copula-based rainfall modeling with the Black-Scholes model for determining insurance premiums. The findings reveal that premium prices depend on rainfall index values, where higher rainfall percentages correspond to higher premium costs.
COMPARISON OF MISSING VALUE IMPUTATION USING MEAN, BAYESIAN KNN, AND NON-BAYESIAN KNN ON TEP GENE EXPRESSION DATA Mastika, Mastika; Siswantining, Titin; Bustamam, Alhadi
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.61-72

Abstract

Analysis of gene expression data, particularly in cancer data, often faces challenges due to the presence of missing values. One approach to overcome this is data imputation. This study evaluates the performance of three imputation methods, namely mean imputation, K-Nearest Neighbors (KNN), and KNN with Bayesian optimization using Gaussian Process modeling, on Tumor Educated Platelets (TEP) gene expression data. Missing values were introduced using Missing Completely at Random (MCAR) gradually at levels of 5%, 10%, 15%, and up to 60%, and performance was evaluated using three metrics: Mean Absolute Error (MAE), Mean Squared Error (MSE), and Normalized Root Mean Squared Error (NRMSE). The results show that the three methods produce relatively similar performance, with differences in MAE, MSE, and NRMSE values only at a small decimal scale. Although Bayesian Optimization is expected to improve the accuracy of KNN, the resulting improvement on this dataset is not significant. These findings indicate that simple imputation such as the average and KNN-based methods still provide competitive results on TEP data with data characteristics that have 14,020,496 zeros out of a total of 16,512,496 existing values, which is approximately 84.91% of the total data.
ANALYSIS MULTILEVEL SURVIVAL DATA USING COVARIATE-ADJUSTED FRAILTY PROPORTIONAL HAZARDS MODEL Sandelvia, Krismona; Effendie, Adhitya Ronnie
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.73-82

Abstract

Multilevel survival data is time-to-event data with a hierarchical or nested structure. This study aims to model the data using the Covariate-Adjusted Frailty Proportional Hazards method, which is an extension of the Cox proportional hazards model with the addition of random effects (frailty). Parameter estimation is performed using a Bayesian approach via Markov Chain Monte Carlo (MCMC). This method is applied to analyze repeated observations of Chronic Granulomatous Disease (CGD) infections, with frailty represented by the hospital and the patient. The results of the data analysis indicate that both hospital and patient frailty significantly influence the time to infection, with patient frailty having a greater effect. Additionally, the treatment variable rINF-g significantly in reducing the risk of serious infection for CGD patients by 64.44%.
FORECASTING THE CLOSING PRICE OF META STOCKS USING A PULSE FUNCTION INTERVENTION ANALYSIS APPROACH Mardianto, M. Fariz Fadillah; Faizun, Nurin; Nauvaldy, Muhammad; Sediono, Sediono
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.83-92

Abstract

Meta Platforms, Inc. (META), the holding company that owns Facebook, Instagram, and WhatsApp, plays a crucial role in advancing artificial intelligence (AI). In early 2024, CEO Mark Zuckerberg announced an ambitious initiative to develop Artificial General Intelligence (AGI), leading to a significant rise in Meta's stock during the first quarter. Consequently, an analysis using the pulse function intervention method was conducted to model and forecast future data. The study utilized weekly data consisting of 124 training and 7 testing observations, spanning from March 13, 2022, to September 15, 2024. The optimal intervention model determined is ARIMA (0,2,1), with parameters (0,0,1) and an intervention point at t = 99. Predictions for a further 8 periods resulted a MAPE of 9.682003% and an MSE of 2411.771. These findings suggest that investors should consider the influence of Zuckerberg's AGI strategy announcement on stock performance. The post-announcement surge indicates a favorable market reaction, and investors should closely follow the AGI project's development to assess META's long-term potential in the technology sector.
Front Matter Vol. 18 No. 1 2025 Statistika, Media
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Cover and Table of Contents
FORECAST EVALUATION OF ARIMA AND ANFIS FOR INDONESIA'S MONTHLY EXPORT (2009-2024) Septiarini, Tri Wijayanti; Rofiqo, Azidni; Pariyanti, Eka; Abdulmana, Sahidan
MEDIA STATISTIKA Vol 18, No 1 (2025): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/medstat.18.1.93-104

Abstract

Indonesia’s export sector is a key driver of economic growth, contributing significantly to foreign exchange, employment, and industrial development. Accurate forecasting of export trends is crucial for policymakers, economists, and businesses in shaping strategies and reducing risks. This study applies the Autoregressive Integrated Moving Average (ARIMA) model to forecast Indonesia’s monthly export values from January 2014 to August 2024. Dataset has been divided into training (75%) and testing (25%) subsets, and the Box-Jenkins methodology was employed, including stationarity testing, identification via ACF and PACF plots, parameter estimation, and residual diagnostics. The optimal ARIMA(1,1,1) model achieved strong predictive performance in RMSE, MSE, and MAPE. To benchmark classical methods against modern approaches, ARIMA was compared with the Adaptive Neuro-Fuzzy Inference System (ANFIS). Results indicated that ARIMA delivered higher accuracy for this dataset, reaffirming the robustness of traditional models when data characteristics align with their assumptions. It has conducted prior research evaluation via 75%:25% holdout and rolling-actual back test. This research demonstrates that classical time-series models remain highly relevant in the era of artificial intelligence, emphasizing the importance of appropriate model selection in economic forecasting.

Page 1 of 2 | Total Record : 11

 1   2 

Filter by Year

2025 2025


Reset
Filter By Issues
All Issue Vol 18, No 1 (2025): Media Statistika Vol 17, No 2 (2024): Media Statistika Vol 17, No 1 (2024): Media Statistika Vol 16, No 2 (2023): Media Statistika Vol 16, No 1 (2023): Media Statistika Vol 15, No 2 (2022): Media Statistika Vol 15, No 1 (2022): Media Statistika Vol 14, No 2 (2021): Media Statistika Vol 14, No 1 (2021): Media Statistika Vol 13, No 2 (2020): Media Statistika Vol 13, No 1 (2020): Media Statistika Vol 12, No 2 (2019): Media Statistika Vol 12, No 1 (2019): Media Statistika Vol 11, No 2 (2018): Media Statistika Vol 11, No 1 (2018): Media Statistika Vol 10, No 2 (2017): Media Statistika Vol 10, No 1 (2017): Media Statistika Vol 9, No 2 (2016): Media Statistika Vol 9, No 1 (2016): Media Statistika Vol 8, No 2 (2015): Media Statistika Vol 8, No 1 (2015): Media Statistika Vol 7, No 2 (2014): Media Statistika Vol 7, No 1 (2014): Media Statistika Vol 6, No 2 (2013): Media Statistika Vol 6, No 1 (2013): Media Statistika Vol 5, No 2 (2012): Media Statistika Vol 5, No 1 (2012): Media Statistika Vol 4, No 2 (2011): Media Statistika Vol 4, No 1 (2011): Media Statistika Vol 3, No 2 (2010): Media Statistika Vol 3, No 1 (2010): Media Statistika Vol 2, No 2 (2009): Media Statistika Vol 2, No 1 (2009): Media Statistika Vol 1, No 2 (2008): Media Statistika Vol 1, No 1 (2008): Media Statistika More Issue