cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Abdullah
Contact Email
abdialam@gmail.com
Phone
+628127580419
Journal Mail Official
data.science.ins@gmail.com
Editorial Address
Jl. Soebrantas Gg. Jelutung Indah no 49 Tembilahan Indragiri Hilir Riau
Location
Kab. indragiri hilir,
Riau
INDONESIA
Data Science Insights
Published by PT Visi Media Network
ISSN : -     EISSN : 30311268     DOI : https://doi.org/10.63017/jdsi.v3i2
Core Subject : Science, Engineering,
Computer Science & IT Engineering
Data Science Insights, with ISSN 3031-1268 (Online) published by PT Visi Media Network is a journal that publishes Focus & Scope research articles, which include Data Science and Machine Learning; Data Science and AI; Blockchain and Advance Data Science; Cloud computing and Big Data; Business Intelligence and Big Data; Statistical Foundation for Data Science; Probability and Statistics for Data Science; Statistical Inference via Data Science; Big Data and Business Analytics; Statistical Thinking in Business; Data Driven Statistical Methods; Statistical Methods for Spatial Data Analytics; Statistical Techniques for Data Analysis; Data Science in Communication; Information and Communication Technology; Graph Data Management for Social Network Applications; Metadata for Information Management; Information/Data: Organization and Access; Information Science and Electronic Engineering; Big Data and Social Science; Data Communication and Computer Network; ICT & Data Analytics. This journal is published by the PT Visi Media Network, which is published twice a year.
Arjuna Subject : Ilmu Komputer (semua kategori) - Ilmu Komputer (Lain-Lain)
Articles 38 Documents
 1   2   3   4 
Liver Disease Prediction using Decision Tree Algorithm Selly
Data Science Insights Vol. 4 No. 1 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i1.138

Abstract

The liver is one of the important organs in the human body that functions to detoxify or neutralize toxins from everything that enters our body, making the body healthier. The liver can be affected by diseases that can disrupt its function; when liver disease attacks, toxins will spread throughout the body, making it unhealthy. Liver disease is a condition caused by viruses, alcohol, lifestyle factors, and others. According to WHO (World Health Organization) data, nearly 1.2 million people die each year, particularly in Southeast Asia and Africa, due to liver disease. Individuals often do not realize or are late in detecting liver disease, so by the time they are examined, the disease is already severe. Early intervention would be better if symptoms are recognized. Data mining can assist in diagnosing liver disease more easily, especially in helping doctors determine whether a patient suffers from liver disease based on symptoms that closely resemble liver conditions. The diagnosis process for liver disease is carried out through classification, resulting in whether the patient has liver disease or not. This study uses five data mining algorithms: Naïve Bayes, K-Nearest Neighbor (KNN), Decision Tree, Random Forest, and Deep Learning.
A Review on Support Vector Machine Problems and Solutions Hiba Al-Dulaimi; Ku Ruhana Ku-Mahamud
Data Science Insights Vol. 4 No. 1 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i1.217

Abstract

In this paper, a review is presented particularly on Support Vector Machine (SVM) problems, so as to understand these problems and to identify the approaches to solve them. The aim is to organize the main SVM problems in a manner that provides a clear view for the readers. The approaches for solving SVM problems were classified into non-simultaneous and simultaneous approaches based on constraints considered in solving the problems. Algorithms for model selection and feature subset selection are classified into heuristic and non-heuristic approaches. Very promising result can be obtained if the bio-inspired algorithms are simultaneously applied with SVM for classification problem.
The Impact of Rainfall Pattern Dataset Construction on Neural Network Performance for Reservoir Water Level Forecasting Wan Hussain Wan Ishak; Raja Nurul Mardhiah Raja Mohamad
Data Science Insights Vol. 4 No. 1 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i1.223

Abstract

Reservoir water level forecasting is a critical component of effective water resources management, supporting flood mitigation, water supply planning, and sustainable reservoir operation, particularly under increasingly variable rainfall conditions. During periods of heavy rainfall, inaccurate or delayed water level prediction may increase flood risk, while during low rainfall seasons, poor forecasting can compromise water storage and operational efficiency. Artificial Neural Networks (ANNs) have been widely adopted for reservoir water level forecasting due to their capability to model nonlinear rainfall–reservoir relationships. However, existing studies largely focus on algorithm selection or architectural enhancement, with limited attention given to how rainfall data representation and dataset construction influence neural network performance. This study addresses this gap by analysing the impact of rainfall pattern dataset construction on ANN performance for reservoir water level forecasting. The primary aim is to evaluate how different rainfall representations affect predictive accuracy when the learning algorithm and training configuration are held constant. Two rainfall pattern datasets were constructed using the same raw rainfall and reservoir water level data from the Timah Tasoh Reservoir, Malaysia. The first dataset represents a compact abstraction of rainfall behaviour using rainfall change indicators derived from day-to-day observations. The second dataset enriches the feature space by incorporating both rainfall change and rainfall intensity categories for each upstream station. In both datasets, the reservoir water level category serves as the prediction target. Prior to model training, redundancy and conflicting data instances were removed to ensure data consistency. A consistent ANN architecture was employed for both datasets and evaluated using 10-fold cross-validation. Model performance was assessed using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). The experimental results demonstrate that the enriched rainfall pattern dataset achieved significantly lower RMSE and MAE values compared to the compact rainfall change dataset, indicating improved learning capability and generalisation performance. Although the enriched dataset required higher computational effort, the improvement in forecasting accuracy was substantial. The findings highlight that dataset construction plays a decisive role in neural-network-based reservoir water level forecasting.
Classification of Stroke Diseases Using Decision Tree Edward Yo
Data Science Insights Vol. 4 No. 2 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i2.201

Abstract

Stroke, a very virulent disease and the risk of carrying death and the most common disease experienced by almost everyone. This study aims to find out whether patients have stroke by using various machine learning algorithms to determine the highest classification algorithm. Disease data was taken from a kaggle created by Federico Soriano with attribute data such as gender, age, hypertension, heart_disease, avg_glucose_level, bmi, smoking_status, and stroke. For data modeling, the cross-validaiton and hold out methods are used with decision tree, random forest, k-NN and naive bayes algorithms. The cross-validation and hold out accuracy values on the decision tree algorithm were 95.74% and 95.72%, random forest was 95.74% and 92.87%, k-NN was 92.46% and 92.87%, and naïve bayes were 91.75% and 93.21%. Then it was implemented in python using a decision tree algorithm with an accuracy value of 91% and a simple GUI application was formed to make it easier to train and test data to show the appearance of stroke.
Application of Predictive Analytics Using Random Forest for Customer Churn Prediction on the Telco Customer Churn Dataset Amelia Contesa
Data Science Insights Vol. 4 No. 2 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i2.234

Abstract

Predictive analytics is an important approach for supporting data-driven managerial decision-making, particularly in addressing customer churn, which can adversely affect company revenue. This study aims to develop a predictive customer churn model and generate insights that can be used to formulate customer retention strategies. The method employs the Random Forest algorithm implemented through RapidMiner for predictive modeling, while Tableau is used for data visualization and interactive exploration of churn patterns. The dataset used is the Telco Customer Churn dataset, which contains information on customer services, charges, and characteristics.The developed model achieved an accuracy of 79.21%, demonstrating a relatively good ability to identify customers who do not churn, although its ability to detect customers who are likely to churn remains limited. Further analysis indicates that the main factors associated with customer churn are contract type, service quality, monthly charges, and customer tenure. Visualization using Tableau reinforces these findings by showing that customers with month-to-month contracts, higher monthly charges, and shorter tenure have a higher risk of churn.The contribution of this study lies in integrating predictive modeling with interactive visualization to produce more comprehensive and accessible insights. This approach can help companies shift from reactive to proactive strategies for customer retention.
Predictive Analytics for Employee Attrition Prediction Using Decision Tree and Random Forest: A 10-Fold Cross-Validation Study Arianto Tarigan
Data Science Insights Vol. 4 No. 2 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i2.237

Abstract

Employee attrition is a major human-resource challenge because unexpected employee departures can increase recruitment and training costs, disrupt operational continuity, and reduce organizational productivity. This study investigates a machine-learning-based predictive analytics approach for employee attrition prediction using Decision Tree and Random Forest with 10-fold cross-validation. The study uses the IBM HR Analytics Employee Attrition dataset, comprising 1,470 employee records and 35 attributes. Data preprocessing includes target-label assignment, missing-value handling, categorical-to-numerical transformation, and normalization. The models are implemented in RapidMiner and evaluated using accuracy, precision, recall, and confusion matrix analysis. The results show that Random Forest outperforms Decision Tree, achieving an accuracy of 85.44% compared with 83.06% for Decision Tree. Random Forest also obtains higher recall for the Attrition = Yes class, increasing from 8.44% with Decision Tree to 14.35%. Despite the improvement, both models exhibit limited minority-class detection, indicating that class imbalance remains a key challenge. Tableau-based exploratory visualization further indicates that factors including MonthlyIncome, OverTime, JobSatisfaction, and TotalWorkingYears are associated with employee attrition. The findings demonstrate that predictive analytics can support data-driven employee-retention decision-making, while future work should address class imbalance and investigate more advanced classification and explainability techniques.  
Predictive Analytics for House Price Estimation Using Multiple Linear Regression: A Managerial Insight Approach Firman Hidayat
Data Science Insights Vol. 4 No. 2 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i2.240

Abstract

In the era of data-driven decision-making, predictive analytics has become an essential approach for transforming raw business data into strategic managerial insights, particularly in the property sector. This study develops an interpretable predictive analytics model for estimating house prices using Multiple Linear Regression (MLR) based on physical and financial housing attributes. Using a housing dataset of 120 records and six variables processed through RapidMiner and Tableau, the findings reveal that building tax amount (β = 0.995, p < 0.001) and building area (β = 1.826, p = 0.022) are the most influential predictors. The number of bathrooms produced a negative coefficient (β = −1665.021, p = 0.017), indicating multicollinearity, while year completed was statistically insignificant (p = 0.276). The model achieved excellent performance with R² = 0.998, MAE = 2,547, and RMSE = 3,215 IDR. This study integrates regression interpretation, predictive analytics, and managerial insight within a practical analytical framework, contributing to property valuation and data-driven decision-making in real estate management.
Comparison of k-Means and Hierarchical Clustering (Ward) Methods for Clustering Regencies/Cities Based on the Food Security Index in West Java Province Al Aghni Naufalia; Nurmawati Ainun Hidayana; Muhammad Bahaudin; Fathir Naufal Hasan; M Al Haris; Alwan Fadlurohman
Data Science Insights Vol. 4 No. 2 (2026): Journal of Data Science Insights
Publisher : PT. Visi Media Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63017/jdsi.v4i2.241

Abstract

Food security in West Java Province exhibits significant variation across regencies/cities, influenced by differences in socioeconomic conditions, health, and sanitation. This condition necessitates an analytical approach based on regional clustering to support the formulation of more precise policies. This study aims to cluster regions based on the Food Security Index (FSI) and to evaluate the performance of K-Means and Hierarchical Clustering (Ward) in producing an optimal cluster structure. The data used comprised 27 regencies/cities with eight standardized FSI indicators. Cluster quality was evaluated using the Silhouette Coefficient and the Davies-Bouldin Index. The results indicate that the optimal FSI structure is divided into three clusters, each exhibiting distinct regional characteristics. Both methods produced consistent patterns; however, K-Means demonstrated superior performance to Ward, yielding a Silhouette Coefficient of 0.3048 and a Davies-Bouldin Index of 1.0725, which were respectively higher and lower than those obtained by Ward (0.2806 and 1.1241). This finding indicates the superiority of K-Means in forming more compact and well-separated clusters within relatively homogeneous data. Further analysis reveals that disparities in food security are primarily influenced by access to sanitation, poverty, education, and the ratio of health workers. This study provides empirical evidence on the effectiveness of clustering methods for FSI analysis and offers a regional clustering framework capable of supporting data-driven food security policy formulation at the regional level.

Page 4 of 4 | Total Record : 38

 1   2   3   4 

Filter by Year

2023 2026


Reset
Filter By Issues
All Issue Vol. 4 No. 2 (2026): Journal of Data Science Insights Vol. 4 No. 1 (2026): Journal of Data Science Insights Vol. 3 No. 2 (2025): Journal of Data Science Insights Vol. 3 No. 1 (2025): Journal of Data Science Insights Vol. 2 No. 2 (2024): Journal of Data Science Insights Vol. 2 No. 1 (2024): Journal of Data Science Insights Vol. 1 No. 1 (2023): Journal of Data Science Insights