Claim Missing Document
Check
Articles

Found 19 Documents
Search

Customer Transaction Clustering with K-Prototype Algorithm Using Euclidean-Hamming Distance and Elbow Method Kuswardana, Dendy Arizki; Prasetya, Dwi Arman; Trimono, Trimono; Diyasa, I Gede Susrama Mas; Awang, Wan Suryani Wan
International Journal of Advances in Data and Information Systems Vol. 6 No. 2 (2025): August 2025 - International Journal of Advances in Data and Information Systems
Publisher : Indonesian Scientific Journal

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59395/ijadis.v6i2.1381

Abstract

This study aims to cluster customer transactions in a Japanese food stall using the K-Prototype Algorithm with a combination of Euclidean-Hamming Distance and the Elbow method. Facing intense industry competition, this study seeks to understand customer purchasing behavior to increase loyalty and sales. From 9.721 initial entries, 9.705 cleaned and transformed records were analyzed. K-Prototype was chosen because of its ability to handle numeric features (Total Sales, Product Quantity) and categorical features (Payment Method, Order Type, Day Category and Time Category). The combination of Euclidean-Hamming distances was used for distance measurement. The optimal number of clusters was determined using the Elbow method, with the results recommending three clusters as the most optimal number. A Silhouette score of 0.6191 indicates a Good Structure clustering result, effectively identifying three distinct customer grouping: "Loyal Regulars" (49.5%), "Casual Shoppers" (42.3%), and "Premium Shoppers" (8.2%). Statistical validity was also tested using ANOVA and Chi-Square, the results showed significant differences between the clusters in numerical and categorical variables with a p-value <0.0001. The clusters are statistically valid in both numerical and categorical aspects. These insights provide an understanding of customer characteristics and reveal a strategically valuable cluster for targeted marketing.
Pendekatan Time Series Decomposition (STL) Dalam Prediksi Kecelakaan Berbasis Kepadatan Lalu Lintas Sebagai Dasar Kebijakan Di Tol Surabaya-Gempol Rizky Mahendra, Rakha; Damaliana, Aviolla Terza; Diyasa, I Gede Susrama Mas
Jurnal Impresi Indonesia Vol. 4 No. 5 (2025): Indonesian Impression Journal (JII)
Publisher : Riviera Publishing

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58344/jii.v4i5.6491

Abstract

Kecelakaan lalu lintas di jalan tol tetap menjadi masalah kritis yang mempengaruhi keselamatan publik dan stabilitasekonomi. Penelitian ini mengusulkan penggunaan dekomposisi Seasonal-Trend menggunakan LOESS (STL) untukmemprediksi risiko kecelakaan berdasarkan data volume lalu lintas di jalan tol Surabaya-Gempol. Data dari Januari 2022hingga Desember 2023, termasuk volume lalu lintas harian dan laporan kecelakaan, diuraikan menjadi komponen tren,musiman, dan residu untuk mengidentifikasi pola. Korelasi positif sedang (r = 0,4882) ditemukan antara volume lalulintas dan frekuensi kecelakaan. Analisis STL mengungkapkan puncak musiman mingguan yang konsisten di akhir pekan,terutama hari Sabtu. Model prediktif yang dikembangkan berhasil mengidentifikasi 11 hari berisiko tinggi pada Januari2024. Berdasarkan temuan tersebut, delapan rekomendasi kebijakan berbasis waktu dirumuskan, termasuk manajemenlalu lintas dinamis, pemantauan real-time, dan peningkatan pengawasan selama periode puncak. Penelitian ini menyumbangkan kerangka kerja berbasis data baru untuk manajemen keselamatan lalu lintas, menggabungkandekomposisi deret waktu dengan panduan kebijakan yang dapat ditindaklanjuti. Tidak seperti penelitian sebelumnya yanghanya berfokus pada prediksi volume, atau pada konteks jalan non-tol, penelitian ini memajukan penerapan STL untukidentifikasi risiko real-time di jalan tol Indonesia. Implikasinya menekankan integrasi sistem lalu lintas cerdas dan potensiprakiraan berbasis STL sebagai fondasi strategi keselamatan jalan nasional.
Classification of Road Damage in Sidoarjo Using CNN Based on Inception Resnet-V2 Architecture Zahrah, Fathima; Diyasa, I Gede Susrama Mas; Saputra, Wahyu Syaifullah Jauharis
Signal and Image Processing Letters Vol 7, No 1 (2025)
Publisher : Association for Scientific Computing Electrical and Engineering (ASCEE)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31763/simple.v7i1.123

Abstract

Road damage is a serious issue in Sidoarjo Regency, posing risks to road users' safety. This study aims to classify road surface conditions using a Convolutional Neural Network (CNN) model based on the Inception ResNet-V2 architecture. The research develops an image-based classification model by combining secondary data from Kaggle and primary data obtained through Google Street View API scraping, along with training strategies such as data augmentation, class balancing, early stopping, and model checkpointing. A total of 885 images were used, categorized into three classes: potholes, cracks, and undamaged roads. The model was trained over 20 epochs with early stopping triggered at epoch 15, when validation accuracy reached 95.95%. Evaluation on the test set showed a test accuracy of 83%. The undamaged road class achieved the highest performance with an F1-score of 0.89, while the pothole class recorded an F1-score of 0.79. The lowest performance was observed in the cracked road class, with an F1-score of 0.65, indicating the model's limited ability to detect fine crack features. This limitation is likely due to class imbalance and visual similarity between classes. Although the model demonstrated good generalization for the two majority classes, the performance gap between validation and test accuracy highlights the need to improve detection for minority classes. Future work is recommended to explore advanced augmentation techniques, increase the representation of minority class data, and consider alternative architectures or ensemble methods to enhance the model’s sensitivity to subtle road damage features.
Food Price Prediction Using the Vector Moving Average (VMA) Model in Surabaya and Malang Najma P., Safira; Trimono; Diyasa, I Gede Susrama Mas
bit-Tech Vol. 8 No. 1 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i1.2428

Abstract

Price fluctuations of animal-based food are a significant issue in Indonesia, especially for low-income communities. In 2023, chicken prices increased by 4.55% and beef prices rose by 11%, contributing to inflation in East Java. Fluctuations in the prices of tuna and milkfish also affected purchasing power and the consumption of animal protein, which remains relatively low compared to other Asian countries. This study aims to predict the prices of animal-based food commodities in Surabaya City and Malang Regency using the Vector Moving Average (VMA) method, which is capable of capturing strong correlations among variables in multivariate time series data. The study covers daily prices per kilogram of beef, chicken, tuna, and milkfish throughout 2023. The 14-day price prediction at the beginning of 2024 shows that the best model for Surabaya is VMA(2), while for Malang Regency, it is VMA(3), selected based on ACF and PACF plots, low AIC values, MAPE values, and consistency of prediction results. The evaluation results using the Mean Absolute Percentage Error (MAPE) indicate that in Surabaya, beef (0.63%), chicken (2.51%), and tuna (4.76%) achieved high prediction accuracy, while milkfish (13.38%) falls into the “good” category. In Malang Regency, the VMA(3) model yielded more consistent prediction results, with all commodities showing MAPE values below 10%: beef (5.62%), chicken (2.43%), tuna (5.61%), and milkfish (2.18%). These results show that the VMA model performs well in capturing the price dynamics of food commodities, as evidenced by the low MAPE values.
Utility-Based Buffer Management for Enhancing DTN Emergency Alert Dissemination in Jakarta's Urban Rail Systems Agussalim, Agussalim; Viet Ha, Nguyen; Putra, Handie Pramana; Adila, Ma’ratul; Diyasa, I Gede Susrama Mas; Rahmat, Basuki
Jurnal Teknik Informatika (Jutif) Vol. 6 No. 4 (2025): JUTIF Volume 6, Number 4, Agustus 2025
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2025.6.4.5241

Abstract

The efficiency of emergency alert dissemination in highly populated and densely urban transport networks, such as Jakarta's integrated rail system, is undermined by sporadic connectivity and limited network resources. In this environment, an initial comparison of baseline Delay-Tolerant Network (DTN) routing protocols revealed that flooding-based routers, such as Epidemic, while achieving above-average delivery rates, suffered from high overhead and poor buffer utilization. This paper fills this gap by proposing the Combined Utility Router, a novel buffer management policy that overcomes the limitations of naive strategies, such as Drop-Oldest. Our approach holistically evaluates a message's value by assigning a weighted utility function based on its Time-To-Live (TTL), estimated total replicas, message size, and a user-defined priority. The router maintains high-value messages by discarding the message deemed the lowest utility score under the buffer constraint. Utility-based simulations in The ONE simulator demonstrate that applying our approach to Epidemic routing improves delivery probability, reduces average latency in high network congestion scenarios, while maintaining overhead rates. This work confirms that, in the context of developing reliable and efficient emergency communication systems for challenging urban topographies, optimizing buffer management extends beyond simply selecting the appropriate protocol.
Daily Forecasting for Antam's Certified Gold Bullion Prices in 2018-2020 using Polynomial Regression and Double Exponential Smoothing Fahrudin, Tresna Maulana; Riyantoko, Prismahardi Aji; Hindrayani, Kartika Maulida; Diyasa, I Gede Susrama Mas
Journal of International Conference Proceedings Vol 3, No 4 (2020): Proceedings of the 8th International Conference of Project Management (ICPM) Mal
Publisher : AIBPM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32535/jicp.v3i4.1009

Abstract

Gold investment is currently a trend in society, especially the millennial generation. Gold investment for the younger generation is an advantage for the future. Gold bullion is often used as a promising investment, on other hand, the digital gold is available which it is stored online on the gold trading platform. However, any investment certainly has risks, and the price of gold bullion fluctuates from day to day. People who invest in gold hopes to benefit from the initial purchase price even if they must wait up to five years. The problem is how they can notice the best time to sell and buy gold. Therefore, this research proposes a forecasting approach based on time series data and the selling of gold bullion prices per gram in Indonesia. The experiment reported that Holt’s double exponential smoothing provided better forecasting performance than polynomial regression. Holt’s double exponential smoothing reached the minimum of Mean Absolute Percentage Error (MAPE) 0.056% in the training set, 0.047% in one-step testing, and 0.898% in multi-step testing.
Comparison of Elbow and Silhouette Methods in Optimizing K-Prototype Clustering for Customer Transactions Kuswardana, Dendy Arizki; Prasetya, Dwi Arman; Trimono, Trimono; Diyasa, I Gede Susrama Mas
EDUTIC Vol 12, No 1: 2025
Publisher : Universitas Trunojoyo Madura

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21107/edutic.v12i1.29744

Abstract

This research presents a comparative analysis of the Elbow and Silhouette methods to identify the ideal number of clusters in applying the K-Prototypes algorithm for customer grouping using purchase transaction data. The K-Prototypes algorithm is employed due to its ability to handle both numerical and categorical data simultaneously. Customer purchase transaction data from the Point of Sale (POS) system is analyzed through preprocessing, feature transformation, and attribute segmentation stages before being clustered using the K-Prototypes algorithm. To identify the optimal cluster count, this study employs two methods: the Elbow and the Silhouette method. The results indicate that the Elbow method produces 2 clusters with a model evaluation score of 0.6368, while the Silhouette method suggests 2 clusters with a slightly lower score of 0.6186. In terms of computational efficiency, the Elbow method also demonstrates a faster processing time results highlight the significance of choosing an appropriate method for identifying the ideal number of clusters, ensuring it aligns with the specific goals of the analysis, whether emphasizing superior inter-cluster distinction or favoring a more parsimonious model configuration.
Implementation Of Natural Language Processing for Spam Email Detection in Outcome Based Education (OBE) Application Diyasa, I Gede Susrama Mas
IJEBD (International Journal of Entrepreneurship and Business Development) Vol 6 No 6 (2023): November 2023
Publisher : LPPM of NAROTAMA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29138/ijebd.v6i6.2587

Abstract

The Natural Language Processing (NLP) approach has been proven to be effective in spam detection in e-mail because of its ability to process text and identify patterns and distinctive characteristics of spam e-mail. Methods in this NLP approach include data pre-processing, such as removing punctuation, irrelevant common words, tokenization, stemming, and others, as well as classification techniques such as Support Vector Classifier (SVC), Naive Bayes, and others. In testing various models, there is one model that shows the highest precision with the number 0.98. This study shows that the NLP approach provides better performance in spam detection compared to other methods. However, it is necessary to improve technology and develop more complex detection methods to improve the performance and accuracy of the email spam detection model
Implementasi K-Means Clustering dengan Menggunakan Data Transaksi Penjualan untuk Penentuan Reward pada Agen Aqua dan Gas LPG FF Tirta Sari, Refika Ayuna; Seta, Henki Bayu; Diyasa, I Gede Susrama Mas
Informatik : Jurnal Ilmu Komputer Vol 18 No 3 (2022): Desember 2022
Publisher : Fakultas Ilmu Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52958/iftk.v18i3.4673

Abstract

Target pelanggan bagi Agen Aqua dan Gas LPG sangatlah penting karena persaingan antar perusahaan sejenis mungkin terjadi dan mengakibatkan hilangnya kepuasan pelanggan. Penelitian ini memiliki tujuan yaitu membangun sistem yang dapat mengelompokkan dan mengetahui tingkat target pelanggan berdasarkan transaksi pembelian agar pihak agen dapat mempertahankan pelanggan yang potensial tersebut dengan cara memberikan memberikan hadiah berupa promo spesial (reward). Hasil penelitian ini ada pengelompokan pelanggan yang mendapatkan reward dan tidak dapat dilihat perbulannya, yaitu pada bulan Januari ada 75 pelanggan mendapatkan reward dan 100 tidak mendapatkan reward, Februari ada 70 pelanggan mendapatkan reward dan 106 tidak mendapatkan reward, Maret ada 80 pelanggan mendapatkan reward dan 96 tidak mendapatkan reward, April ada 35 pelanggan mendapatkan reward dan 141 tidak mendapatkan reward, Mei ada 65 pelanggan yang mendapatkan reward dan 111 tidak yang mendapatkan reward, Juni ada 43 pelanggan yang mendapatkan reward dan 133 tidak mendapatkan reward l, Juli ada 77 pelanggan yang mendapatkan reward dan 99 tidak mendapatkan reward, Agustus ada 48 pelanggan yang mendapatkan reward dan 128 tidak mendapatkan reward, September ada 94 pelanggan mendapatkan reward dan 82 tidak mendapatkan reward, dan Oktober ada 94 pelanggan mendapatkan reward dan 82 tidak mendapatkan reward.Kata Kunci: Clustering, K-Means Clustering, Reward, Davies-Bouldin Index