Setijohatmo, Urip Teguh
Unknown Affiliation

Published : 4 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 4 Documents
Search

Development of the Shortest Path Navigation Feature in a 360° Virtual Campus Tour Using Dijkstra's Algorithm Alifi, Muhammad Riza; Hodijah, Ade; Setijohatmo, Urip Teguh; Wulan, Sri Ratna; Hayati, Hashri
Journal of Artificial Intelligence and Software Engineering Vol 5, No 2 (2025): June
Publisher : Politeknik Negeri Lhokseumawe

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30811/jaise.v5i2.6839

Abstract

A 360° virtual campus tour allows users to independently explore all available scenes in the form of 360° panoramic photos through a self-guided navigation feature. However, not all navigation tools provided are capable of generating route recommendations for users to follow. This presents a challenge, as users may feel overwhelmed when deciding where to begin and end the tour—particularly when the number of scenes reaches into the hundreds. In certain scenarios, prolonged interaction within a virtual reality environment may lead to discomfort due to motion sickness. Implementing a shortest path algorithm offers a potential solution by guiding users through recommended routes, thereby improving exploration efficiency and reducing interaction time. This study integrates a shortest path-based navigation feature into a virtual campus tour using Dijkstra’s algorithm, consisting of: (1) a front-end navigation component for the user interface of route searching, and (2) a back-end routing component that processes pathfinding using a graph-based structure. The implemented navigation feature demonstrates high efficiency, with an average execution time of only 4.94 ms and low memory consumption, as measured by a resident set size of 710.47 KB and used heap memory of 668.61 KB.
Enhancing News Similarity with Chunking Strategy and Hyperparameter Setting on Hybrid SBERT - Node2Vec Model Permadi Supriyo, Reza Ananta; Setijohatmo, Urip Teguh; Maspupah, Asri
Journal of Information System and Informatics Vol 7 No 3 (2025): September
Publisher : Universitas Bina Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51519/journalisi.v7i3.1180

Abstract

The proliferation of online news necessitates accurate article similarity systems to combat information overload, yet models based solely on semantic content often ignore crucial structural context like news source and publication date. This research proposes and evaluates a hybrid embedding model that integrates semantic representations from Sentence-BERT (SBERT) with structural representations from Node2Vec. A series of quantitative experiments were conducted on the challenging, multilingual SPICED dataset to determine the optimal model configuration. Using Mean Squared Error (MSE) for evaluation, the results show that a per-paragraph chunking strategy yielded the best performance. This strategy's effectiveness was validated by the identical performance of an optimal fixed-size chunk (450 characters with a 64 overlap), a value that aligns closely with the dataset's average paragraph length. Furthermore, a community-focused (BFS-like) Node2Vec configuration (p=1.0, q=2.0, l=60) was identified as optimal for the structural component. Significantly, the final hybrid model (MSE = 0.1434) proved superior to both the purely semantic (MSE = 0.1449) and purely structural models (MSE = 0.2512). This study concludes that the fusion of content and context provides the most comprehensive and accurate representation for news similarity detection.
Development of a Scientific Article Recommendation Web Application Using a Hybrid Recommender System: A Case Study in Computer Science Az-zahra, Adinda Raisa; Prawira, Alfien Sukma; Nugraha, Mahesya Setia; Hodijah, Ade; Setijohatmo, Urip Teguh; Setiarini , Siti Dwi; Wisnuadhi , Bambang
Media Jurnal Informatika Vol 17 No 2 (2025): Media Jurnal Informatika
Publisher : Teknik Informatika Universitas Suryakancana Cianjur

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35194/mji.v17i2.5672

Abstract

The rapid increase in scientific publications has created significant challenges for researchers in finding relevant literature. Conventional citation-based recommender applications often have drawbacks, such as bias toward popular articles and vulnerability to manipulation through citation cartels, which reduce objectivity. To address these limitations, this development aimed to design and develop a web-based scientific article recommendation application using a hybrid recommender system approach. The development followed the waterfall methodology, covering requirements analysis, design, implementation, and testing stages. The hybrid approach combines Content-based filtering by analyzing content similarity and Collaborative filtering based on user interaction history. Scientific articles and user preferences were modeled in a graph database to map relationships, with the implementation of Graph Data Science Library using algorithms named K-Nearest Neighbor, Degree centrality, and PageRank.  Based on 101 black-box unit test cases, the application successfully delivered three main recommendation features by integrating content analysis—based on access history and currently viewed articles—with user preference modeling through peer institutions. The testing results confirm that all recommendation functions operated as intended across various user scenarios.  Overall, the developed application provides multiple recommendation features that enhance objectivity and relevance, supporting researchers, students, and practitioners in discovering
Application of Named Entity Recognition (NER) in Job Vacancy Matching Using an Ontology-Based Approach (Case Study: Information Technology Sector) Gunawan , Rizki; Hodijah, Ade; Taqwim , Muhammad Ikhsan Maulana; Siti Ababil , Afyar; Setijohatmo, Urip Teguh; Wulan, Sri Ratna; Alifi , Muhammad Riza; Sari , Aprianti Nanda; Hayati, Hashri
Media Jurnal Informatika Vol 17 No 2 (2025): Media Jurnal Informatika
Publisher : Teknik Informatika Universitas Suryakancana Cianjur

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35194/mji.v17i2.5675

Abstract

The dissemination of job vacancies through online platforms still faces limitations in understanding the semantic relationships between the skills possessed by job seekers and the qualifications required by a job position. This mismatch results in an inefficient search process and longer search times. This study aims to develop a semantic-based job vacancy recommendation application (talent matching) using a skill ontology approach. One of the main challenges in developing the ontology is the lack of standardized data structures in job vacancy postings, particularly in the job description section. To address this issue, Named Entity Recognition (NER) techniques are applied to automatically extract skill entities from job description texts. The extracted results are then classified into a taxonomy structure using SkillsGPT, thereby forming a hierarchical skill concept model semantically represented within the ontology using Protégé. The matching process between user skills and job qualifications is conducted through semantic similarity calculations employing the Sánchez Similarity method. Job vacancy data are collected via web scraping, while system development follows the Rational Unified Process (RUP) methodology and is evaluated using Black Box testing. Evaluation results demonstrate that the developed system is capable of providing semantically relevant job vacancy recommendations according to the user's skill profile. Therefore, this study contributes both theoretically and practically to the development of ontology-based recommendation systems, particularly in the automated modeling of skill taxonomies from unstructured data.