cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Dr. Suci Astutik, S.Si,. M.Si.
Contact Email
suci_sp@ub.ac.id
Phone
+6281334404567
Journal Mail Official
jasds.ub@ub.ac.id
Editorial Address
Jl. Veteran, Malang 65145, East Java, Indonesia
Location
Kota malang,
Jawa timur
INDONESIA
JASDS: Journal of Applied Statistics and Data Science
Published by Universitas Brawijaya
ISSN : -     EISSN : 30484391     DOI : https://doi.org/10.21776/ub.jasds
Core Subject : Science, Education,
Computer Science & IT Mathematics
JASDS : Journal of Applied Statistics and Data Science (e-ISSN: 3048-4391) is a journal managed by Universitas Brawijaya , Malang, Indonesia, and associated with FORSTAT (Forum Pendidikan Tinggi Statistika) which is published twice a year (in March and October). The objectives of Journal of Applied Statistics and Data Science are to publish and disseminate high quality of original research papers about the application of statistics and data science in many areas, or case driven theoretical development of statistics and data sciences. The journal covers the following topics: Experimental Design, General Linear Model and Generalized Linear Model, Bayesian, Time Series, Spatial, Econometrics, Big Data, Machine Learning, Panel Model, Computational Statistics, Operation Research, Actuarial and Finance, Statistical Quality Control, and related topics. Upon its submission, the Editor in Chief decides on the suitability of the paper’s content for the aim and scope of JASDS. If the Editor in Chief considers the paper is suitable, then the paper will be sent for peer reviewing by two peer reviewers. Journal of Applied Statistics and Data Science maintains double anonymity, so neither the peer reviewers nor the author(s) can be identified by one another. The peer reviewers are the respectful scholars of the areas.
Arjuna Subject : Matematika - Statistik dan Probabilitas
Articles 31 Documents
 1   2   3   4 
Optimizer Performance Test on CNN Long Short-Term Memory Network for Car Sales Forecasting
Journal of Applied Statistics and Data Science Vol. 2 No. 1 (2025): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2025.002.01.6

Abstract

In the automotive industry, forecasting future demand is particularly crucial due to the complexity of production processes and supply chains. This article examines the comparative performance of a hybrid CNN-LSTM model for car sales forecasting, utilizing seven optimization algorithms: Adam, RMSprop, SGD, Adagrad, Adadelta, Adamax, and Nadam. Each optimization method has its own advantages. For instance, Adam offers fast convergence, while RMSprop is more effective in handling large gradient fluctuations. Adagrad is well-suited for managing gradient magnitude variations, whereas Adadelta addresses Adagrad’s limitations. Adamax is ideal for models with a broader parameter space, and Nadam combines Nesterov Accelerated Gradient and Adam, making it suitable for tasks requiring both momentum and adaptive learning. This study demonstrates that the CNN-LSTM model optimized with Nadam delivers the best performance, achieving a Mean Squared Error (MSE) of 35,383.14 and a Root Mean Squared Error (RMSE) of 188.10. In comparison, traditional methods such as ARIMA yield an MSE of 59,105.94 and an RMSE of 243.11. These findings indicate that the CNN-LSTM model optimized with Nadam outperforms conventional time series forecasting methods in predictive accuracy.
Discriminant Analysis Consumer Behavior in Selecting Online Health Service Applications (Telemedicine) (Study of Mathematics and Natural Sciences Faculticity Brawijaya University Bachelor Student)
Journal of Applied Statistics and Data Science Vol. 3 No. 1 (2026): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2026.003.01.2

Abstract

The research was conducted to examine the consumer behavior characteristics, particularly among undergraduate students of the Faculty of Mathematics and Natural Sciences (FMIPA) at Universitas Brawijaya, in selecting online healthcare service applications (telemedicine) using quadratic discriminant analysis. The data used was primary data, collected through quota sampling technique via a survey by distributing questionnaires to 100 respondents. The dependent variable in this study was the selection of applications, divided into Halodoc, Alodokter, and KlikDokter, with the independent variables being trust, ease of use, and price. The analysis was applied to obtain the classification function, classification results, and the strongest discriminating variable. The analysis results showed that the strongest discriminating variable in the formed discriminant function was price, with the largest standardized coefficient among trust and ease of use. The classification accuracy obtained was 58%.
An Estimation of Missing Value on Time Series Data Using ARMA Interpolation, Average Value and Kalman Filtration on Boeing Co Company Stock Price Data
Journal of Applied Statistics and Data Science Vol. 2 No. 2 (2025): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2025.002.02.3

Abstract

Missing data in time series data become a problem because it causes the time series to decrease. The small time series causes problems estimating model parameters, so missing data must be imputed with a value. This study aims to compare and discuss the best methods for estimating missing data, several methods that can be used to predict missing data on stock price adjustments of closed Boeing Company, which consist of the ARMA (Autoregressive Moving Average) Interpolation Method, the Kalman Filtering Method, and the Average Value Method. MAPE (Mean Absolute Percentage Error) as a measure of the goodness of the estimator to the actual value is used in determining the best imputation method among the three methods, the results of the ARMA Interpolation Method using the ARMA (1,0) time series model produce a MAPE value of 2.52%, the Kalman Filtering Method of 3.15% and Method Average Value of 5.30%. The ARMA imputation method is the best for estimating missing data on closing stock price time series data with Boeing Co adjustments, with the smallest MAPE value compared to the Kalman Filtering Method and the Average Value Method of 2.52%, which means the imputation is very good.
Comparison of Opinion Mining Algorithms (Decision Tree and Naïve Bayes) for iOS 16 Opinion Classification on Twitter
Journal of Applied Statistics and Data Science Vol. 2 No. 2 (2025): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2025.002.02.4

Abstract

This study analyzes public response to the release of iOS 16 by Apple Inc. on September 13, 2022. The iOS operating system, exclusive to iPhone devices, generated diverse opinions across social media platforms, particularly Twitter. The research aims to compare the effectiveness of Decision Tree and Naïve Bayes algorithms in classifying public sentiment toward iOS 16, considering the simplicity of both algorithms as an advantage in classification method implementation. The research methodology involved data collection using web crawling techniques through Jupyter Notebook, resulting in 14,946 Indonesian-language tweets gathered during the period of September 13-27, 2022. The tweets underwent a series of processing stages, including comprehensive text preprocessing and sentiment analysis using VADER lexicon polarity detection to categorize tweets into positive, negative, and neutral sentiments. The dataset was divided with a 90:10 ratio for training and testing data. The findings demonstrate that the Decision Tree algorithm outperforms Naïve Bayes in classifying opinions about iOS 16. The Decision Tree achieved superior performance metrics with a Recall of 79.39%, Precision of 79.73%, and Accuracy of 80.60%.
Geographically Weighted Regression (GWR) Modeling with Adaptive and Fixed Kernel Functions on Stunting Event in South Sulawesi Province
Journal of Applied Statistics and Data Science Vol. 3 No. 1 (2026): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2026.003.01.4

Abstract

The issue of stunting is currently a national priority in Indonesia. South Sulawesi Province has a fairly high prevalence of stunting and is included in the top 10 provinces with the highest stunting rates in Indonesia. So research is needed to determine the variables that influence the prevalence of stunting to support prevention efforts. This research discusses Geographically Weighted Regression (GWR) modeling using fixed kernel and adaptive kernel weighting functions to analyze variables that influence stunting cases in South Sulawesi Province. The independent variables used are the percentage of low birth weight (LBW), children who do not have IMD, adequate sanitation, children who do not have MCH books, and poor people, while the dependent variable is the prevalence of stunting. The data used is secondary data for 2023 sourced from publications by the Central Statistics Agency (BPS). The GWR method is applied to capture the influence of spatial heterogeneity between regions using six weighting functions: Fixed Gaussian Kernel, Fixed Bisquare Kernel, Fixed Tricube Kernel, Adaptive Gaussian Kernel, Adaptive Bisquare Kernel, and Adaptive Tricube Kernel. The selection of the best model is carried out based on the Akaike Information Criterion (AIC). The research results show that the Adaptive Bisquare Kernel weighting function provides the best results with the smallest AIC value. The variables that have a significant effect on the prevalence of stunting are the percentage of children without IMD, adequate sanitation and poor people.
Estimation and Forecasting of HPG's Return Rate in Vietnam Using The ARIMA-GARCH Model
Journal of Applied Statistics and Data Science Vol. 2 No. 2 (2025): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2025.002.02.5

Abstract

The volatility of stock indices in the stock market is a crucial basis for market evaluation and trend prediction, often having a significant impact on investment and trading decisions. Therefore, analyzing and forecasting these changes helps to effectively manage risk and optimize returns for investors. This study analyzes the volatility of the return rate of HPG stock based on its closing prices from 24/09/2021 to 24/09/2024. The most suitable model selected for forecasting is ARIMA(11,0,19)-GARCH(1,1). The results indicate that, between 25/09/2024 and 04/10/2024, the return rate of HPG stock is expected to experience slight fluctuations with varying positive and negative returns. However, the decline is not significant and does not substantially affect the long-term trend of the stock.
Forecasting the Price of Broiler Eggs Using the Generalized Space-Time Autoregressive Method in 3 Regions with The Highest Egg Prices in East Java
Journal of Applied Statistics and Data Science Vol. 3 No. 1 (2026): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2026.003.01.1

Abstract

Time series is the data that is compiled based on the time sequence. Not only time series data, but there is also some data that have a relationship between time and location, which is called space-time data. One of the analyses used in space-time data is Generalized Space-Time Autoregressive (GSTAR). This study applied the GSTAR method to create a model for the price of broiler eggs in Pamekasan Regency, Sampang Regency, and Sumenep Regency and also do the forecasting for the price of broiler eggs for the next 12 months. Modeling using GSTAR for the price of broiler eggs shows that all parameters are not significant. Therefore, changes in the prices of broiler eggs are not influenced by the price of broiler eggs from other regencies or prices from their regency. However, based on the MAPE and RMSE, the best model is GSTAR ( )I(1) with inverse distance weight. The RMSE and MAPE of the GSTAR ( )I(1) model with inverse distance weight are 1.565,85 and 4,68%. Keywords: broiler egg price, GSTAR, inverse distance weight, uniform weight, normalized cross-correlation weight.
Forecasting Dengue Hemorrhagic Fever (DHF) Cases In Semarang City Using Machine Learning Extreme Gradient Boosting (XGBoost)
Journal of Applied Statistics and Data Science Vol. 2 No. 2 (2025): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2025.002.02.6

Abstract

Dengue Hemorrhagic Fever (DHF) remains a significant health issue in Semarang City, with the number of cases fluctuating from year to year. To support appropriate policy-making, an accurate forecasting method is required. This study aims to forecast the number of DBD cases based on monthly data from January 2021 to March 2025 using the Extreme Gradient Boosting (XGBoost) algorithm. Data was obtained from the Semarang City Health Department and analyzed based on the gender of the patients. The XGBoost model was chosen for its ability to capture complex patterns in time series data. Model evaluation using the MAE and RMSE metrics showed satisfactory results, with an MAE value of 7.87 and an RMSE of 8.83 for males, and an MAE of 3.50 and an RMSE of 4.42 for females. Forecast results for the period from April to August 2025 indicate that DBD cases among males are likely to remain stable at around 6–7 cases per month, while cases among females are expected to remain steady at approximately 9 cases. These findings suggest that XGBoost is effective for forecasting DBD cases and can serve as a tool for future health policy planning.
Clustering PT Indosat Tbk B2B Customers Using K-Medoids Algorithm with One Hot Encoding Technique
Journal of Applied Statistics and Data Science Vol. 3 No. 1 (2026): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2026.003.01.6

Abstract

Cluster analysis is a method for grouping objects based on characteristic similarities. The K-medoids algorithm, an extension of K-means, offers superior robustness against outliers by utilizing representative objects (medoids). This study analyzes transaction data from Indosat Ooredoo Hutchison’s B2B directorate Malang branch for the second semester of 2024. The objective is to segment B2B customers to assist the directorate in formulating targeted business strategies. Using K-medoids with Gower distance, 81 companies were categorized into five clusters: High-Tech Aggressive Potential (18 companies) with the highest purchasing power and active digitalization; Churn Potential (13 companies) at risk of service termination; High-IT Potential (25 companies) focusing on IT digitalization; Stable Potential (9 companies) showing consistent purchasing behavior; and Low-Tech Moderate Potential (16 companies) with moderate demand for complex technology. These findings enable the B2B directorate to implement data-driven strategies, such as personalized retention programs and tailored technological offerings. This segmentation provides a strategic foundation for optimizing customer relationship management and resource allocation.
Latin Hypercube Design Under Second and Third-Order Models: A Prediction Variance Approach
Journal of Applied Statistics and Data Science Vol. 3 No. 1 (2026): Journal of Applied Statistics and Data Science
Publisher : Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21776/ub.jasds.2026.003.01.3

Abstract

Prediction variance describes the error involved with making a prediction using a response surface model. This study examines the prediction variance performance of Latin Hypercube Designs (LHDs) within second- and third-order response surface models. G-optimality, I-optimality criteria, and Fraction of Design Space (FDS) plots were employed to assess the predictive capabilities and accuracy of LHDs. The findings reveal that LHDs perform better under third-order models when evaluated using the G-optimality criterion, while under the I-optimality criterion, LHDs perform better in second-order models. The FDS plots further indicate that as the number of factors increases, the prediction errors across models become approximately similar.

Page 3 of 4 | Total Record : 31

 1   2   3   4 

Filter by Year

2024 2026


Reset
Filter By Issues
All Issue Vol. 3 No. 1 (2026): Journal of Applied Statistics and Data Science Vol. 2 No. 2 (2025): Journal of Applied Statistics and Data Science Vol. 2 No. 1 (2025): Journal of Applied Statistics and Data Science Vol. 1 No. 2 (2024): Journal of Applied Statistics and Data Science Vol. 1 No. 1 (2024): Journal of Applied Statistics and Data Science