Journal of Informatics, Electrical and Electronics Engineering

Klasterisasi Provinsi Di Indonesia Berdasarkan Indeks Masyarakat Digital Tahun 2025 Menggunakan Algoritma K-Means Clustering

2026-08-23T18:13:01+00:00

Abstract-Digital transformation is one of Indonesia's national development priorities, yet the level of digital readiness across regions still shows a considerable gap. The 2025 Digital Society Index (Indeks Masyarakat Digital or IMD) measures each province's digitalization achievement through four pillars, namely Infrastructure and Ecosystem, Digital Literacy, Empowerment, and Employment. This study aims to group 38 Indonesian provinces based on these four pillars using the K-Means Clustering algorithm so that the pattern of regional digital disparity can be identified objectively. The dataset used is secondary data of the 2025 IMD, which was standardized using the Z score method prior to clustering. The optimal number of clusters was determined through a combination of the Elbow method and the Silhouette Coefficient, leading to the selection of three clusters with a Silhouette Coefficient of 0.262. The clustering results show that the provinces are divided into three groups, namely a digitally advanced cluster consisting of 18 provinces with an average IMD of 49.18, a developing cluster of 13 provinces with an average IMD of 43.16, and a lagging cluster of 7 provinces with an average IMD of 35.11 that is dominated by the Papua region. The Empowerment pillar was found to be the sharpest differentiator between clusters, while the Employment pillar was relatively uniform across provinces. These findings are expected to serve as a consideration for policy makers in designing more targeted digital transformation equity programs.

Penerapan Metode Profile Matching Dalam Menentukan Produk Bedak Padat Untuk Jenis Kulit Berminyak Berdasarkan Konsumen

2026-10-20T13:00:11+00:00

Every woman desires to look beautiful and possess healthy, radiant skin, while also seeking solutions for various skin concerns. Pressed powder is a cosmetic product that can help address such issues; it is designed to conceal facial imperfections, including those related to oily skin. Consequently, selecting the right pressed powder is crucial for managing and controlling facial oil, thereby reducing excess oiliness. Profile Matching is an algorithm used in Decision Support Systems (DSS) to compare the profile values ??of specific criteria (competency values) against the profile values ??of various alternatives. This method calculates a "GAP" (representing the alignment between the required and actual values); a lower GAP value indicates a better fit. By applying the Profile Matching method—utilizing data gathered via questionnaires distributed through social media—six alternatives were evaluated: A1 (Emina), A2 (Luxcrime), A3 (Marcks), A4 (Skintific), A5 (Somethinc), and A6 (Wardah). Six criteria were considered: C1 (Price), C2 (Packaging), C3 (Ingredients), C4 (Durability), C5 (Shade), and C6 (Composition). The analysis identified A6 (Wardah) as the best pressed powder product, with a score of 46.68. The Profile Matching method serves as a valuable tool for consumers, assisting them in selecting the most suitable pressed powder for their specific skin type and ensuring they make an informed choice.

Implementasi Metode Forward Chaining pada Sistem Pakar Penentuan Jenis Kulit Berbasis Web

2025-12-25T13:12:47+00:00

Determining your facial skin type independently is often a challenge for people due to a lack of knowledge about skin characteristics and limited access to dermatology experts. Misdiagnosing your skin type can lead to choosing inappropriate skincare products. Therefore, a system capable of assisting in the rapid and systematic process of diagnosing skin types is needed. This research developed a rule-based expert system for determining skin types using the forward chaining method. This method works by tracing facts in the form of symptoms entered by the user to conclusions based on a predetermined rule base. This system is designed to utilize parameters such as hydration level, oil content, sensitivity, and skin moisture as the basis for the inference process. System testing was conducted using 50 test data sets obtained from Kaggle. The test results showed that the system was able to classify skin types with an accuracy of 82%, with 41 data sets correctly identified, 2 data sets misclassified, and 7 data sets unable to be classified due to rule limitations. The results of this study demonstrate that the forward chaining method is quite effective in handling the rule-based skin type determination process. The developed system can be an alternative tool for quickly identifying skin types and has the potential for further development in the fields of AI and dermatology.

Overcoming Tree Extrapolation Limits in Gold Price Forecasting: A Stationary Return-Reconstructed XGBoost Approach

2026-10-20T17:32:31+00:00

This study addresses the fundamental tree extrapolation limit inherent in gradient-boosted decision tree ensembles when forecasting non-stationary financial time series during historical market rallies. Standard raw-price regression models fail to project unseen price regimes because decision trees assign static scalar predictions to terminal leaf nodes, capping predictions at historical maximums. To overcome this spatial limitation, we propose a Stationary Return-Reconstructed XGBoost framework that shifts the machine learning target from non-stationary spot prices to stationary daily relative returns before deterministically re-integrating predicted returns into nominal price levels. Evaluated on an out-of-sample testing set spanning 2022 to 2024, the return-reconstructed framework completely eliminates flatline prediction plateaus during record-setting gold price rallies exceeding 2,800 USD per troy ounce. The proposed model achieves an R-squared score of 0.9905, a Root Mean Squared Error of 28.17 USD per troy ounce, and a Mean Absolute Percentage Error of 1.12 percent, outperforming the raw-price baseline model by 92.75 percent in average error reduction. Furthermore, game-theoretic SHAP analysis validates that daily return forecasts are driven by macroeconomically sound transmission channels, predominantly led by the US Dollar Index and monthly historical price memory

Sistem Pendukung Keputusan Pemilihan E-Wallet Terbaik Menggunakan Metode TOPSIS

2026-10-20T15:25:55+00:00

The advancement of technology has led to numerous innovations in financial services, commonly known as financial technology (fintech), which have replaced the role of cash in non-cash payment transactions. This has prompted many electronic money providers in Indonesia to utilize smartphones to enhance their services through digital wallet applications. E-wallets allow users to conduct non-cash transactions to purchase goods or services, such as money transfers, bill payments, buying credit, or merchant payments. Some users may find the usage of digital wallet systems confusing, and even those who understand may still seek further benefits from choosing the best E-Wallet based on criteria such as security, feature availability, user convenience, transaction fees, and merchant integration. Therefore, the calculation from the above research utilizes the Technique for Orders Preference by Similarity to Ideal Solution (TOPSIS) method with Rank Order Centroid (ROC) weighting, involving 6 alternatives and 5 criteria, resulting in a higher-ranked decision value. The top-ranked E-Wallet application is Dana (A3), with a preference value of 0.9656