cover
Contact Name
Siti Aminah
Contact Email
sitiaminah@ubhinus.ac.id
Phone
+62341-560823
Journal Mail Official
lppm@ubhinus.ac.id
Editorial Address
Jl. Raya Tidar No 100 Malang
Location
Kota malang,
Jawa timur
INDONESIA
Smatika Jurnal : STIKI Informatika Jurnal
ISSN : 20870256     EISSN : 25806939     DOI : https://doi.org/10.32664/smatika
Core Subject : Science,
SMATIKA: STIKI Informatika Jurnal is a journal published by Lembaga Penelitian & Pengabdian kepada Masyarakat (LPPM) of Universitas Bhinneka Nusantara Malang. The scope of this journal in the field of Computer Science, Information Systems, and Information Management.
Articles 307 Documents
Development of a CNN-Based Household Waste Image Classification Model for Automatic Identification Achmad isra; Azmuri Wahyu Azinar; Suhendro Busono; Hamzah Setiawan
Smatika Jurnal : STIKI Informatika Jurnal Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM Universitas Bhinneka Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v16i03.2436

Abstract

The problem of household waste in Indonesia remains a major challenge due to the low level of waste sorting and the lack of technology utilization in the waste identification process. This study aims to develop a Deep Learning-based inorganic waste identification system using the Convolutional Neural Network (CNN) method to improve the accuracy of automatic classification in the household environment. The research method used is an experiment with a quantitative approach. The dataset used consists of 92 images of inorganic waste covering the categories of plastic, paper, glass, and metal. The data went through a preprocessing stage in the form of image resizing to 224 × 224 pixels, normalization, and data augmentation, then divided into training and testing data with a ratio of 80:20. The CNN model was trained using variations in epoch parameters, batch size, and learning rate, then evaluated using metrics of accuracy, precision, recall, F1-score, and confusion matrix. The results showed that the training parameters had a significant effect on model performance. The optimal value was obtained at epoch 20 with 97% accuracy, a batch size of 8 with 97% accuracy, and a learning rate of 1×10⁻³ with 90% accuracy. The best parameter combination produced a model with 97% classification accuracy, along with balanced precision, recall, and F1-score values. These results demonstrate the effectiveness of the CNN method in automatically identifying inorganic waste types. The developed system has the potential to be applied in household environments to assist waste sorting, improve waste management efficiency, and support environmental conservation by raising public awareness of the importance of sustainable waste management. Although the model achieved 97% accuracy under a single 80:20 train-test split, its performance decreased substantially to 32.69% ± 6.93% under stratified 5-fold cross-validation, indicating potential overfitting and limited generalization.
Segmentation of Hikers Preferences for Hiking Boot Recommendations Using K-Means and SAW Dian Fitri Islamiati Munisah; Devita Maulina Putri
Smatika Jurnal : STIKI Informatika Jurnal Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM Universitas Bhinneka Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v16i03.2468

Abstract

The increasing number of mountain climbers in Indonesia, particularly beginners, has led to issues in selecting the right hiking shoes. Without data-driven guidance, shoe selection tends to be influenced solely by brand popularity rather than suitability with physical needs and trail characteristics, which potentially increases the risk of injuries such as blisters or sprains. This problem is exacerbated by the lack of a system capable of accommodating differences in preferences among climber segments, considering that the needs of technical trail climbers differ significantly from those doing light hikes (day hikes).This research aims to identify mountain climbers' preference patterns and build a decision support system for hiking shoe recommendations based on user segmentation. The approach used is the integration of the K-Means Clustering method to group climbers based on shoe attribute preferences, and the Simple Additive Weighting (SAW) method to generate objective brand recommendation rankings based on weights extracted from the clustering results. Data was collected from 154 climber respondents on Mount Buthak and Mount Penanggungan, East Java, using a ranking-based preference questionnaire.The clustering results with the best performing though still suboptimal, configuration of K=2 (Silhouette Score 0.4506) produced two segments: Cluster 0, which prioritizes comfort and durability (80 respondents), and Cluster 1, which is oriented towards Brand and shoe Weight (74 respondents). The application of SAW yielded different recommendations for each segment, with Consina being the top recommendation for Cluster 0 (score 0.8663) and Nineten for Cluster 1 (score 0.8810). These results prove that the integration of K-Means and SAW can produce more objective and personalized recommendations compared to approaches based solely on brand popularity, while also serving as a foundation for market segmentation strategies for outdoor equipment manufacturers.
Cluster Analysis of Indonesia’s Electronic Integrated Circuit Export Destination Using K-Means and K-Medoids Christian Andrew; Rahmatina Hidayati
Smatika Jurnal : STIKI Informatika Jurnal Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM Universitas Bhinneka Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v16i03.2474

Abstract

One of Indonesia’s export commodities is Electronic Integrated Circuits (EICs). Understanding the characteristics of EIC export destination countries is important for identifying export patterns and to support strategic decision-making. Clustering is one approach that can be used to analyze export destination countries by grouping them based on their similar characteristics. The objective of this study is to evaluate the performance of K-Means and K-Medoids algorithms in analyzing Indonesia’s EIC export destination countries. The data used in this study were obtained from Badan Pusat Statistik (BPS) and consist of export value and weight data from 2021 to 2025 containing 84 destination countries. This dataset is selected because limited studies have compared K-Means and K-Medoids on Indonesia’s EIC export destination countries. The data from BPS were then preprocessed, then normalized by applying Min-Max normalization method. The Elbow Method was subsequently employed to find the optimal number of clusters, with the analysis indicating that four clusters were optimal. Following the Elbow Method, K-Means and K-Medoids algorithms were then applied and assessed based on the Davies-Bouldin Index and Silhouette Score. The findings of this study demonstrate that K-Means outperformed K-Medoids in both the DBI and Silhouette Score by achieving lower DBI value and higher Silhouette Score. The results of both algorithms revealed that most destination countries belonged to the very low export value and weight cluster, accounting for around 88% of countries in K-Means and 60% in K-Medoids. The findings suggest that K-Means is more suitable for clustering Indonesia’s EIC export destination countries and can provide useful insight into export patterns that may support future trade analysis and decision-making.
Development of an Web-Based Order Management System for Tailoring Businesses with Production Monitoring and WhatsApp Notifications Muhammad Afwa Nawafal; Indyah Hartami Santi; Udkhiati Mawaddah
Smatika Jurnal : STIKI Informatika Jurnal Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM Universitas Bhinneka Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v16i03.2488

Abstract

Tailoring businesses, such as Anita Modiste Boutique, often face operational bottlenecks due to reliance on manual record-keeping for order processing and production monitoring, which leads to data inconsistencies, communication delays, and inefficient decision-making. This study proposes an integrated digital workflow model to modernize these operations, developed through the Rapid Application Development (RAD) methodology to ensure system alignment with administrative needs. The research contributes a novel approach by integrating customer order management, down payment tracking, and real-time internal coordination via an automated WhatsApp Gateway, serving as an active coordination mechanism between production staff and the boutique owner. The system’s functionality was verified with an 84.62% success rate in initial Black Box testing, followed by 100% compliance after iterative refinements, while its acceptance was empirically validated via User Acceptance Testing (UAT) involving 61 respondents, resulting in a User Satisfaction Index (USI) score of 89.44% ("Excellent"). These findings demonstrate that digitizing internal coordination, beyond mere customer-facing administration, is critical for enhancing the operational performance and transparency of MSMEs, providing a functional and user-accepted digital solution for tailoring businesses.
Automatic Menu Recommendation for Industrial Catering Using Constraint-Based Filtering, K-Means Clustering, and Apriori Association Rule Mining Elang Rezky Pratama; Yulian Findawati; Cindy Taurusta; Yunianita Rahmawati
Smatika Jurnal : STIKI Informatika Jurnal Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM Universitas Bhinneka Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v16i03.2498

Abstract

Industrial catering services that supply daily meals to factory workers face recurring operational challenges, particularly in planning menus that are balanced in terms of cost, ingredient availability, and nutritional variety. Manual menu planning is time-consuming and often results in inconsistent decisions, especially when ingredient stocks fluctuate daily. This study presents the design and implementation of an automatic daily menu recommendation system for Puteri Catering, an industrial catering service in Sidoarjo, East Java, Indonesia. The system integrates three complementary approaches: Constraint-Based Filtering to validate menu candidates against current ingredient stock and budget limits, K-Means Clustering to categorize menus into three cost groups (Cheap, Moderate, and Expensive) based on their production cost per serving (HPP), and Apriori Association Rule Mining to discover frequent menu combination patterns from historical transaction data. The system was developed using the Laravel 11 framework with a MySQL database, while the machine learning components were implemented in Python using the Scikit-learn and MLxtend libraries. The dataset used consists of 43 menu items, 62 raw ingredients, and 60 historical daily transactions. Evaluation results show that K-Means clustering achieved a Silhouette Score of 0.5794, indicating a good separation between clusters. The Apriori algorithm successfully generated 83 association rules, with the strongest rule being Rawon → Mendol with a confidence of 88.9% and a lift of 3.56. Functional testing using the Black Box method across 47 test cases covering 9 system modules yielded a 100% success rate. The system successfully generates three menu package recommendations per day, each with automatic ingredient stock deduction, providing a practical and data-driven solution for daily operational planning in industrial catering services.
Comparative Performance Analysis of MobileNetV2, ResNet50V2, and EfficientNetB0 Models Using Transfer Learning for Medicinal Plant Leaf Image Classification Mohammad Yoga Firnanda; Rohman Dijaya; Ika Ratna Indra Astutik; Yulian Findawati
Smatika Jurnal : STIKI Informatika Jurnal Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM Universitas Bhinneka Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v16i03.2520

Abstract

Accurate identification of medicinal plant leaves supports the correct and safe use of herbal remedies, yet manual identification is often unreliable because many species share closely related leaf shapes, textures, and colors. Transfer learning has emerged as a practical strategy for image classification tasks of this kind, since it reuses the feature representations of pre-trained convolutional neural networks (CNNs) and thereby lowers both training time and computational demand. This study aims to compare the performance of three transfer learning backbones—MobileNetV2, ResNet50V2, and EfficientNetB0—on a five-class medicinal plant leaf recognition task covering guava (Psidium guajava), basil (Ocimum basilicum), aloe vera (Aloe vera), betel (Piper betle), and pandan (Pandanus amaryllifolius), using a working dataset of 1,750 leaf images (350 images per class before deduplication), and to statistically verify whether the performance differences among the three architectures are significant. Data were partitioned 80:10:10 for training, validation, and testing, and each architecture was trained across five independent runs under an identical configuration; Accuracy, Precision, Recall, and F1-score were used to quantify performance, followed by the Shapiro–Wilk and Kruskal–Wallis tests. MobileNetV2 produced the strongest mean Accuracy (98.27% ± 0.52), while ResNet50V2 led on Recall. The Kruskal–Wallis results, however, showed no statistically significant gap between the three architectures on any metric (p > 0.05). In other words, MobileNetV2, ResNet50V2, and EfficientNetB0 can be regarded as functionally equivalent for this classification task, and the practical choice among them should therefore weigh factors such as inference speed, model size,  and  deployment  constraints  rather  than  average predictive performance alone.
White-Box Basis Path Testing of the Bellman-Ford Algorithm in a Package Delivery Application Abdi Pandu Kusuma; Ayuna Rahim Imannur
Smatika Jurnal : STIKI Informatika Jurnal Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM Universitas Bhinneka Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v16i03.2576

Abstract

Route determination applications based on the Bellman-Ford algorithm are widely adopted because they handle weighted graphs iteratively, including negative weights. However, testing that only checks output correctness, such as black-box testing, cannot guarantee that every execution path, particularly inside the nested loops of the edge-relaxation process, behaves as intended, so hidden logical errors may remain undetected. This study verifies the structural correctness of the nearest delivery-distance module of a web-based package delivery application that implements the Bellman-Ford algorithm, using white-box testing with the basis path technique: flowgraph construction, cyclomatic complexity calculation, independent path determination, and the design and execution of test cases. The flowgraph consists of 15 nodes, 18 edges, and 4 decision nodes, giving a cyclomatic complexity V(G) of 5, which is classified as simple and low-risk; five independent paths cumulatively cover all 18 edges. Executing the five path-based test cases, together with one supplementary test on a larger sample graph, produced outputs identical to the expected results in every case, so the implementation is structurally valid within the scope of the basis path coverage criterion. Nested loops do not by themselves raise V(G): the number of independent paths depends on the number of decision points rather than on the number of iterations, so V(G) should not be read as a proxy for the O(|V||E|) running time of Bellman-Ford. Path-based testing additionally reached three paths, namely an empty edge set, negative-weight cycle detection, and an inner loop that is never entered, which the earlier black-box testing had not exercised. The five-path test set provides a replicable structural testing procedure for similar graph algorithm implementations, to be complemented by loop and boundary testing.

Filter by Year

2015 2026


Filter By Issues
All Issue Vol 16 No 03 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 16 No 02 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 16 No 01 (2026): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 15 No 02 (2025): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 15 No 01 (2025): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 14 No 02 (2024): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 14 No 01 (2024): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 13 No 02 (2023): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 13 No 01 (2023): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 12 No 02 (2022): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 12 No 01 (2022): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 11 No 02 (2021): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 11 No 01 (2021): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 10 No 02 (2020): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 10 No 01 (2020): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 9 No 02 (2019): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 9 No 01 (2019): SMATIKA Jurnal : STIKI Informatika Jurnal Vol 8 No 02 (2018): SMATIKA Jurnal STIKI Informatika Jurnal Vol 8 No 01 (2018): SMATIKA Jurnal STIKI Informatika Jurnal Vol 7 No 02 (2017): Smatika Jurnal : STIKI Informatika Jurnal Vol 7 No 01 (2017): Smatika Jurnal : STIKI Informatika Jurnal Vol 6 No 02 (2016): Smatika Jurnal : STIKI Informatika Jurnal Vol 6 No 01 (2016): Smatika Jurnal : STIKI Informatika Jurnal Vol 5 No 02 (2015): Smatika Jurnal : STIKI Informatika Jurnal More Issue