JITSI : Jurnal Ilmiah Teknologi Sistem Informasi

Evaluasi Penggunaan Fitur Voice Chat Pada Aplikasi Discord Berdasarkan Perspektif User Persona

2025-10-09T03:15:42+01:00

This study evaluates the usability level of the Voice Chat feature on the Discord application using the System Usability Scale (SUS) approach from a user persona perspective. The main objective of this study is to understand how usability perceptions differ across user groups based on their specific characteristics, needs, and constraints. A quantitative method was used by distributing the SUS questionnaire to 60 respondents consisting of students, teachers, and private workers. The results showed an average SUS score of 62.375 which placed the Discord Voice Chat feature in the "Marginal High" (Grade D) category, indicating a level of usability that still needs to be improved. Analysis based on user personas revealed significant differences in user experience: students were more familiar with Discord but faced technical constraints such as microphone interference, while teachers and private workers had difficulty in navigation and role management. These findings emphasize the importance of a more inclusive and responsive feature design to the needs of users from various backgrounds. This study provides insights for the development of Discord that is oriented towards a more personalized and contextual user experience.

Analisis User Experience Pada Aplikasi Mobile Legend Bang Bang Menggunakan Metode Game Experience Questionnaire (GEQ)

2025-10-14T04:19:42+01:00

Mobile legend bang-bang is an online game application with 5VS MOBA background, at some time the Mobile legend bang-bang application received negative reviews from mobile legend players related to the bad community, related to the latest mobile legend update, this study aims to determine the user experience after using the mobile legend bang-bang application regarding the level of satisfaction from the user experience with the game experience questionnaire method, this method has 3 modules but the module used by researchers is post game, namely the module that measures the level of user satisfaction after playing the mobile legend bang-bang game. The results obtained in this study regarding the bang-bang mobile legend using the game experience method and the post game module get the results that negative experience, Tiredness and returning to reality get low results where the lower the TCR value obtained, the better if the value is low, while the positive experience gets a pleasant value that is in good results, from the results that have been obtained show that the analysis of user satisfaction using post games on the game experience questionnaire gets good results from the respondents that feel satisfied and feel good after playing mobile legend bang-bang based on the data that has been analyzed..

Sentiment Analysis of Pinterest Application User Reviews Using ANN, CNN, and RNN Methods

2025-10-09T03:15:42+01:00

The changes occurring in the pinterest application have sparked numerous opinions expressed on google playstore, both positive and negative. The purpose of this study is to analyze the sentiment of Indonesian public towards the Pinterest application through user reviews on the platform. The research method employed in this study is qualitative, utilizing data collection techniques through scraping user reviews and interviews. The research was conducted from October 2024 to January 2025. The data used consists of 2000 reviews collected in the years 2023 and 2024. This research uses 3 deep learning methods because they can understand large amounts of data. Others prefer machine learning as their research method because it is easier and less complicated. The RNN method is an effective method for performing sentiment analysis with large amounts of data. This is supported by research results indicating that the RNN (Recurrent Neural Networks) method achieved the highest accuracy in sentiment analysis, reaching 65.17%, followed by two other deep learning methods, namely CNN (Convolutional Neural Networks) and ANN (Artificial Neural Networks). The RNN method is effective because it is supported by high precision and recall values. The author suggests that future research should explore other methods and expand data from different platforms to gain a broader perspective.

Implementasi Aplikasi Pengelolaan Pengajuan Peminjaman Melalui Agen pada PT Adira Pandaan

2025-10-09T03:15:42+01:00

This study aims to improve the efficiency and accuracy of the loan application process at PT. Adira Dinamika Multi Finance Pandaan through the development of a cloud-based mobile application. The system was developed using the Agile Scrum methodology, leveraging Flutter and Firebase technologies integrated with QR Code functionality and automated notifications via WhatsApp Bot. The research involves five key user roles: customer, agent, registration admin, submission admin, and supervisor. Evaluation was conducted through black box testing, usability testing based on the Likert scale, and eligibility classification testing using fuzzy logic. The results indicate that all system features function as expected, with high usability levels and positive user satisfaction. The implemented fuzzy logic model successfully classified loan eligibility with an accuracy rate of 92%. The system has been proven to accelerate the application process, reduce manual errors, and enhance transparency and operational efficiency. Therefore, the system is deemed suitable for full-scale implementation to support the digitalization of loan application processes within the company.

Implementasi Algoritma Yolo Untuk Mendeteksi Jalan Berlubang dan Retak

2025-10-09T03:15:42+01:00

Roads are transportation facilities intended for traffic as well as public infrastructure that supports the mobility and accessibility of road users. Road damage is commonly found and is often identified manually. Since the invention of Computer Vision in the 1950s, object detection and classification have attracted significant interest in various sectors, including industry and medicine. Since then, many studies have been conducted using various Deep Learning algorithms capable of detecting objects. This research utilizes the YOLOv8 (You Only Look Once) algorithm to detect and classify images. This algorithm works by predicting bounding boxes and class probabilities on the entire image in a single capture. In its application, the data is divided into training, validation, and testing datasets, consisting of images of damaged roads categorized into cracks and potholes. By applying specific configurations in the YOLOv8 algorithm, the resulting output includes the Confusion Matrix calculations. The research involves analyzing results from various data train and validation splits: 70%-30%, 80%-20%, and 90%-10%, with training epochs of 25x, 50x, and 100x, to evaluate how training iterations affect performance outcomes. The results indicate that the highest confidence level is achieved when using a 90%-10% split between training and validation data, reaching 97% confidence, with mAP of 93.2%, F1-Score of 88.7%, Recall of 90.8%, and Precision of 86.7% at the 100th epoch.

Analyzing Determinants of E-Government Use Among Students: An Integrated UTAUT and E-Government Adoption Model

2025-10-09T03:15:42+01:00

Digitalization in the education sector, particularly through e-government, is key to enhancing service effectiveness and efficiency in academic information systems. However, the application of this technology often faces challenges in terms of student acceptance, as it is the primary user. Using the Unified Theory of adoption and Use of Technology (UTAUT) and the E-Government Adoption approach, this study endeavors to explore the determinants affecting the adoption and utilization of the Academic Information System at Tidar University. The variables studied include performance expectations, convenience expectations, social influence, supporting conditions, security and privacy risks, and trust in government. This study used quantitative methods and involved 247 active students of the State Administration Study Program as respondents. Questionnaires were utilized for data collection in this research, and the responses were examined using Structural Equation Modeling (SEM) with a Partial Least Squares (PLS) method. The analysis showed that the utilization of the system was substantially influenced by variables including performance expectancy, social influence, trust in governmental institutions, facilitating conditions, and behavioral intention. Meanwhile, effort expectancy, perceived risk, and security risk have no significant effect. These results verify that the optimal utilization of academic information systems on campus is significantly influenced by performance expectation, social influence, and trust in government

Prediksi Harga Laptop Berdasarkan Spesifikasi Menggunakan Algoritma K-Nearest Neighbour

2025-10-09T03:15:42+01:00

The development of information technology is driving the increasing need for laptops as the primary computing device for various groups. Laptop prices vary greatly, influenced by specifications such as processor type, RAM capacity, storage, and additional features. This research aims to develop a laptop price prediction system based on specifications using the K-Nearest Neighbor (K-NN) algorithm, integrated into a web-based application with a Laravel API backend and MySQL database. The K-NN algorithm works by finding data with the most similar characteristics to determine the price estimate. The laptop price dataset used underwent preprocessing, including converting categorical data to numerical data so it could be processed by the model. Testing was conducted with varying K values of 3, 5, 7, and 9, using the metrics of accuracy, Mean Absolute Error (MAE), and Root Mean Square Error (RMSE). The test results showed that K = 5 provided the best performance with an accuracy of 0.89, MAE of 0.15, and RMSE of 0.18. Too small a K value makes the model sensitive, while too large a K value makes the model too general, making K = 5 the optimal point. This model is capable of producing predictions with a low error rate and high accuracy, proving the effectiveness of K-NN in predicting laptop prices based on specifications such as processor brand, processor type, RAM, storage capacity, and screen size. The success of this research also proves that the integration of Flutter, Laravel API, and MySQL can support the development of a responsive, interactive, and beneficial laptop price prediction system for both potential buyers and sellers

Implementation of Query Optimization Methods in a Web-Based Point of Sale System at Toko Eci Dolok

2025-10-09T03:15:42+01:00

Penelitian ini mengkaji penerapan optimasi kueri pada sistem Point of Sale (POS) berbasis web di Toko Eci Dolok. Permasalahan utama yang teridentifikasi adalah pencatatan transaksi secara manual yang tidak efisien dan menyulitkan penyusunan laporan penjualan. Tujuan penelitian ini adalah merancang sistem POS yang lebih cepat dan efisien melalui optimasi kueri dengan menggunakan operasi seleksi, proyeksi, dan join skeleton. Data yang digunakan terdiri dari 772 item produk, transaksi penjualan, kasir, dan pemasok. Hasil percobaan menunjukkan bahwa optimasi kueri dapat mengurangi total biaya eksekusi dari 60.036 halaman menjadi 20.012 halaman, menghasilkan penghematan sekitar 66,7% dan peningkatan kecepatan eksekusi tiga kali lipat. Kesimpulannya, metode ini efektif dalam meningkatkan efisiensi pemrosesan data pada sistem POS dan berpotensi untuk diaplikasikan dalam skala yang lebih besar, seperti e-commerce atau sistem informasi manajemen.

Virtual Reality Development for Computer Assembly Practical Simulation Learning Media

2025-10-09T03:15:42+01:00

Computers function as a central instrument in today’s learning processes; consequently, students are often expected to possess baseline competencies in assembling them. This study develops a Virtual Reality (VR)–based computer-assembly practicum simulation using Unity and Oculus Quest 2 to help students at Amikom University Yogyakarta grasp assembly procedures in an interactive, realistic way. The application follows the Multimedia Development Life Cycle (MDLC), which comprises concept, design, material collection, assembly, testing, and distribution. Core functionality includes a sequenced simulation of assembly steps—installing the motherboard, processor, RAM—and a concluding functionality test. Evaluation employed Blackbox Testing to verify system behavior and the System Usability Scale (SUS) to assess usability. Blackbox Testing showed that all features worked properly. The mean SUS score was 74.63, categorized as Grade B, indicating good usability and ease of use. Overall, the findings underscore VR’s promise as an effective and innovative medium for computer-assembly practicums

Results of Teachers' Frequency of Use of PHET Sımulatıon and Analysıs

2025-10-09T03:15:42+01:00

This article examines the use of PhET interactive simulations in biology classes at higher and secondary schools in Azerbaijan, focusing on the frequency of use by teachers and the impact on learning outcomes for learners. The study surveyed 100 teachers of various subjects. A mixed-methods approach was used, including observations and a comparative analysis of student performance before and after the integration of PhET simulations. Research shows that the majority of teachers (72%) use PhET simulations occasionally or regularly in their lessons. The introduction of these digital tools has resulted in significant improvements in students’ comprehension and mastery of complex biological topics. Specifically, student achievement in subjects such as mitosis, meiosis, DNA structure, and photosynthesis has increased by 31% to 42%. The author presents examples supported by detailed tables and graphical analyses. The article also offers a comparative discussion of PhET simulations compared to other digital learning tools and highlights their unique advantages. It is recommended that teachers be encouraged to use digital tools through professional development courses and to promote the prospects of using interactive simulations. For this, it has become necessary to update curricula and expand virtual laboratory resources. Overall, PhET simulations increase interactivity and student motivation in biology teaching, making learning more effective. These findings emphasize the importance of integrating innovative digital technologies into education and offer a promising future for their widespread application

Analysis of Acceptance of STARS Website Usage Using TAM Method: Case Study of Satya Wacana Christian University Students

2025-10-09T03:15:42+01:00

In the ever-evolving digital era, students' ability to utilise an effective academic information system is important to optimally support the learning process and management of education administration. This study aims to explore how students accept and use the STARS website through the Technology Acceptance Model (TAM) approach. The research method used in this study is a quantitative approach by collecting data through questionnaires distributed to Satya Wacana Christian University students who use the STARS website. Data analysis was carried out using inferential statistical techniques to evaluate the effect of Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Attitude Toward Using (ATU) and Behavioural Intention to Use (BI) on Actual System to Use. The results showed that Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) have a significant influence on the attitude (ATU) of students towards using STARS. In addition, Behavioural Intention to Use (BI) was found to be a strong predictor of actual use of the system, indicating that the more positive students' attitude towards STARS, the higher their intention to use it in academic activities. These findings underscore the importance of paying attention to perceived ease and benefits in the development of academic systems to increase the effectiveness of technology use in educational environments.

Deep Learning Neural Dalam Analisis Sentimen : Sebuah Studi Literatur

2025-10-09T03:15:42+01:00

This study provides a thorough review of the literature on the application of deep learning in sentiment categorization. The major purpose is to collect statistical data on deep learning research for sentiment analysis and hybrid model construction. The research discovered that the most often utilized deep learning algorithms in 2022-2024 were BERT, LSTM, and GRU, each with varied degrees of accuracy. Specifically, GRU had the highest accuracy (98%), followed by LSTM (93.58%) and BERT (91.37%). In addition, 31% of the examined publications modified these methods to create new hybrid models. Among them, the RoBERTA and LSTM hybrid models achieved the highest accuracy (91.01%). This systematic review examines the changing landscape of sentiment analysis using deep learning, focusing on the efficacy of hybrid models in boosting classification accuracy.