Scientific Journal of Information System

Comparative Analysis of Cloud Service Models for Professional Use: IaaS, PaaS, and SaaS

2026-03-30T06:19:53+00:00

This study aims to conduct a structured comparative analysis of cloud computing service models-Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS)-for professional use across multiple sectors. A quantitative comparative approach was employed using data collected from scientific literature and semi-structured interviews involving 15 professionals from education, business, and technology sectors. Each model was evaluated based on five parameters: flexibility, scalability, cost efficiency, user control, and sector relevance using a Likert scale (1–5). The results indicate that IaaS achieved the highest score in flexibility (5.0) and user control (5.0), PaaS showed balanced performance across development-related parameters (average score 4.2), while SaaS demonstrated the highest cost efficiency (5.0). These findings highlight that no single model is universally superior, and selection should be aligned with organizational priorities. This study contributes by providing a parameter-based quantitative comparison framework to support decision-making in cloud service adoption.

Machine Learning-Based Knowledge Trend Analysis Using LDA for Strategic Decision-Making

2026-04-04T06:21:48+00:00

In the digital economy, Knowledge Management Systems (KMS) often fail to provide actionable insights due to information overload, leaving valuable expertise fragmented and underutilized. This research aims to integrate Machine Learning (ML) to transform passive data into proactive strategic foresight by analyzing knowledge trends. Using a longitudinal dataset of search logs and document metadata, the study implements a text-mining pipeline centered on Latent Dirichlet Allocation (LDA) to extract thematic clusters. The model identified eight distinct knowledge domains, with "Advanced Data Analytics" emerging as a high-growth sector (TVI = +0.13), while a critical "Knowledge Gap" in cybersecurity was detected where search demand outpaced document supply by 58%. This study contributes by proposing a Trend Velocity Index (TVI) to quantify knowledge evolution and detect knowledge gaps, providing a robust framework for leaders to optimize resource allocation and ensure institutional agility.

Machine Learning for Predicting Property Purchase Behavior: A Systematic Literature Review

2026-03-31T05:39:05+00:00

This study aims to examine the application of machine learning algorithms in predicting property purchase behavior based on consumer data. The main problem addressed is the limited use of intelligent data analysis in understanding consumer behavior in the Indonesian property sector, despite increasing market data availability. This research employs a systematic literature review approach by analyzing studies published in the last five years, focusing on classification algorithms such as Decision Tree, Random Forest, and Support Vector Machine (SVM). The analysis includes data collection, evaluation, and synthesis of selected studies. The results indicate that algorithm performance varies depending on data characteristics and application context. Random Forest tends to show strong performance in terms of accuracy and robustness, while Decision Tree and SVM also demonstrate competitive results in certain scenarios. These findings reflect general trends rather than definitive conclusions. Key factors influencing property purchase decisions include location, price, and developer reputation. In conclusion, machine learning has significant potential to support data-driven decision-making in the property sector. Future research should integrate real-time and more diverse data to improve predictive model accuracy

Analysis of Priority-Based Communication Feature Using UI/UX Design Thinking Model: A Case Study of WhatsApp

2026-04-13T09:15:31+00:00

The rapid growth of instant messaging applications has significantly transformed the way individuals communicate in both personal and professional contexts. However, the increasing volume of incoming messages often leads to information overload, making it difficult for users to distinguish between important and less relevant conversations. This study aims to design and implement a priority-based communication feature using a Design Thinking approach, with WhatsApp as a case study. Unlike conventional chronological message ordering, the proposed system allows users to manually define priority contacts through a “High Priority Mode” feature, enabling important conversations to be automatically positioned at the top of the chat list. In addition, the system introduces visual differentiation in notifications to highlight messages from priority contacts. A prototype interface is developed to support intuitive configuration and improve usability.

The results indicate that the proposed feature enhances message visibility, reduces the risk of overlooking important communications, and improves overall user efficiency. This study demonstrates that a user-centered, rule-based approach can provide a practical and effective solution for managing communication priorities in messaging applications.

Implementation of the Naive Bayes Algorithm for Classification of Public Service Complaints in E-Government at Kunciran Indah Tangerang

2026-04-21T03:20:34+00:00

The implementation of e-government at the local government level is essential for improving the quality and efficiency of public services. However, the management of public service complaints at Kelurahan Kunciran Indah, Tangerang, is still conducted manually, leading to delays and inefficiencies in handling citizen reports. This study aims to implement the Naive Bayes algorithm to automatically classify public service complaints within an e-government system. A quantitative computational approach was employed using a dataset of 50 complaint records categorized into four classes: infrastructure, cleanliness, service, and administration. Data preprocessing techniques, including case folding, tokenization, and stopword removal, were applied prior to model training. The Naive Bayes classifier was used to build a classification model and evaluate its performance. The results show that the proposed model achieved an accuracy of 90%, demonstrating good performance in classifying text-based complaints across all categories. This indicates that the Naive Bayes algorithm is effective for supporting automated complaint classification in local government services. The implementation of this system can improve service efficiency, accelerate response time, and assist decision-making processes. Nevertheless, the study is limited by the relatively small dataset, and future research is recommended to utilize larger and more diverse data to enhance model performance.