cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Jamaluddin
Contact Email
hengkitamando26@gmail.com
Phone
+6281397181985
Journal Mail Official
publication.aptikomsumut@gmail.com
Editorial Address
Jl. Alumni No.3, Padang Bulan, Kec. Medan Baru, Kota Medan, Sumatera Utara 20155
Location
Kota medan,
Sumatera utara
INDONESIA
Journal of Computer Science and Research
Published by Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer Provinsi Sumatera Utara
ISSN : -     EISSN : 29862337     DOI : -
Core Subject : Economy, Science, Engineering,
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Library & Information Science
Journal of Computer Science and Research (JoCoSiR) is aimed to publish research articles on theoretical foundations of information and computation, and of practical techniques for their implementation and application in computer systems. Journal of Computer Science and Research (JoCoSiR) published quarterly and is a peer reviewed journal covers the latest and most compelling research of the time. Journal of Computer Science and Research (JoCoSiR) is managed and published by APTIKOM Wilayah 1 Sumatera Utara.
Arjuna Subject : Ilmu Komputer (semua kategori) - Kecerdasan Buatan
Articles 63 Documents
 2   3   4   5   6 
SENTIMENT ANALYSIS OF E-COMMERCE REVIEWS WITH NATURAL LANGUAGE PROCESSING (NLP) Idhami, Rahmat; Saputra, Andri; Fadly, Taufa; Silaban, Robet
Journal of Computer Science and Research (JoCoSiR) Vol. 2 No. 2 (2024): April: Computer Science
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v2i2.61

Abstract

E-commerce in Indonesia is growing rapidly, with Shopee as a leading platform. This study uses Natural Language Processing algorithms to analyze customer satisfaction sentiment from reviews on the Google Play Store. The results identify issues related to courier services and provide recommendations for improving service quality, delivery tracking systems, and overall customer satisfaction and loyalty towards Shopee. This chapter describes the research methodology for sentiment analysis of Shopee reviews using Natural Language Processing methods. These stages include data collection, cleaning, pre-processing, labeling, data separation, classification, and negative word analysis. This study aims to identify the dominant negative sentiment in Google Play Store reviews. This study outlines data scraping, cleaning, pre-processing, labeling, and Natural Language Processing classification to identify negative words in Shopee user reviews. This method provides insights into courier service issues and recommendations for couriers frequently highlighted in reviews, with a focus on future service improvements. Based on the study, Natural Language Processing is effective in identifying positive and negative sentiment in Shopee with an accuracy of 86-87%. Negative sentiment was dominant (62.5%), particularly regarding "recommended couriers," with complaints about delays and unprofessionalism. Recommendations included improving courier service quality, delivery tracking systems, customer communication, and courier training and supervision to improve customer satisfaction.
Implementation of Grey Wolf Optimizer (GWO) Algorithm for Predicting Multidrug Resistance Patterns in Bacteria Harefa, Ade May Luky
Journal of Computer Science and Research (JoCoSiR) Vol. 2 No. 2 (2024): April: Computer Science
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v2i2.62

Abstract

The emergence of multidrug-resistant (MDR) bacterial pathogens poses a critical threat to global health, demanding intelligent and adaptive predictive systems. This study proposes the application of the Grey Wolf Optimizer (GWO) algorithm as an innovative computational approach for predicting and analyzing multidrug resistance patterns in clinical bacterial isolates. Unlike conventional statistical methods that often fail to handle complex, nonlinear biomedical data, GWO effectively balances exploration and exploitation through swarm intelligence inspired by wolf hierarchy and hunting behavior. A dataset of 10,700 clinical bacterial samples obtained from Kaggle was analyzed, encompassing antibiotic susceptibility profiles and clinical parameters such as patient comorbidities and hospitalization history. The data were normalized and optimized using GWO to identify the most influential attributes contributing to antibiotic resistance. Experimental results demonstrate that GWO achieves strong stability in convergence, efficiently identifying dominant resistance predictors such as CTX/CRO, FOX, and IPM. Compared to traditional optimization methods, GWO offers improved accuracy and robustness in feature weighting and selection. The study concludes that GWO provides a scalable and interpretable framework for multidrug resistance prediction, enabling early identification of critical resistance trends. The implementation of this approach can assist healthcare institutions in formulating more precise antimicrobial stewardship strategies and controlling the spread of resistant pathogens in clinical environments.
Analysis and Classification of IT Professions in the Marketplace Using the Support Vector Machine Method: Analysis and Classification of IT Professions in the Marketplace Using the Support Vector Machine Method Ardya, Dwika; Iqbal, Muhammad
Journal of Computer Science and Research (JoCoSiR) Vol. 2 No. 2 (2024): April: Computer Science
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v2i2.64

Abstract

The development of the digital industry in Indonesia has driven an increasing demand for professional workers in the information technology (IT) sector. Along with this, the need arises to understand and map salary levels based on job profiles to create transparency and efficiency in the recruitment process. This study aims to predict the salary categories of IT professionals using the Support Vector Machine (SVM) method in well-known marketplace companies such as Gojek, Shopee, Tokopedia, Traveloka, Tiket.Com and Bukalapak. The dataset used contains 611 data entry records with attributes of company, work location, experience and skills as well as salary. The preprocessing process consists of label encoding, numeric normalization, and multi-hot encoding for skill features. The salary categories are divided into three: low, medium, and high. The SVM model is trained with the Radial Basis Function (RBF) kernel and evaluated with accuracy, precision, recall, and f1-score metrics. The evaluation results show that the SVM model is able to classify salary categories with an accuracy of 82%. This model shows the best performance in the Medium salary category with an f1-score of 0.93. This study proves that SVM can be used as an alternative to build an effective IT Salary Category prediction system.
Model Predictive Analysis of Performance in Training and Course Institutions Using Naive Bayes and K-Means Clustering Eko Budianto; Muhammad Iqbal
Journal of Computer Science and Research (JoCoSiR) Vol. 3 No. 1 (2025): Jan: Computer Science
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v3i1.68

Abstract

The performance of course and training institutions (LKP) is a crucial factor in determining the quality of non-formal education in Indonesia. Performance assessments are currently conducted manually using the National Accreditation Board for Non-Formal Education (BAN-PNF) assessment instrument, which is time-consuming and prone to subjectivity. This research aims to develop a predictive analysis model for the performance of course and training institutions using a combination of the Naive Bayes and K-Means Clustering methods. The K-Means Clustering method is used to group institutions based on similar characteristics across key variables such as trainers, infrastructure, curriculum, management, and graduate outcomes. These clustering results are then used as additional features for the Naive Bayes classification model to predict performance categories (high, medium, or low). Testing of 150 institutions' data showed a predictive accuracy of 89.2%, with three main clusters representing high-, medium-, and low-performing institutions. This model has the potential to become a data-driven tool for governments and institutions to conduct performance evaluations quickly, objectively, and adaptively to changes in training data.
Classification of Student Activity Status Using Machine Learning Algorithms at Royal University Hermawan, Rudi; Muhammad Iqbal
Journal of Computer Science and Research (JoCoSiR) Vol. 3 No. 1 (2025): Jan: Computer Science
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v3i1.69

Abstract

Inactivity is a significant challenge that impacts academic performance, retention rates, and the operational effectiveness of higher education institutions. Royal University faces an urgent need to identify students at risk of becoming inactive early, so that academic interventions can be carried out appropriately and effectively. This study aims to develop a classification model for student inactivity status (Active or Passive) using a machine learning approach, by testing three main algorithms: Decision Tree (DT), Support Vector Machine (SVM), and Random Forest (RF). The dataset used consists of 642 student entries, including academic information such as Grade Point Average (GPA), total credits taken, attendance percentage, number of courses per semester, and semester level. The methodology steps include data cleaning and transformation, splitting the dataset into 80% training data and 20% testing data using a random sampling method ( train_test_split with random_state = 42), model training, and performance evaluation using accuracy, precision, recall, and F1-score metrics. The experimental results show that DT and SVM achieve the highest accuracy of 98.44%, with maximum precision in predicting active students, while RF excels in recall (0.96), making it more effective in detecting active students at risk of being missed. Feature importance analysis reveals that GPA and attendance are the most determining factors in predicting student active status, while the number of courses, credits taken, and semester level have a lower additional influence. The primary contribution of this research is the provision of an accurate and practically applicable classification model, enabling universities to conduct automated student monitoring, proactive academic interventions, and data-driven decision-making. Implementing this model in academic information systems can improve the effectiveness of advising programs, reduce the risk of student inactivity , and support efforts to improve retention and graduate quality. This research also emphasizes the importance of contextual features in improving prediction accuracy and provides insights that can be leveraged for the development of data-driven academic strategies
Analysis of Patient Satisfaction Toward the Implementation of the Bed Management Application at Langsa General Hospital: A Case Study of Bed Management System Deployment JB, Salwa Nur; Fachrurazy, Fachrurazy; Lola Astri Nadita; Sri Hidayati
Journal of Computer Science and Research (JoCoSiR) Vol. 3 No. 2 (2025): April: Artificial Intelligence
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v3i1.66

Abstract

The digital transformation of healthcare has become a strategic imperative for improving hospital efficiency, transparency, and patient-centered service quality. This study examines the impact of the Implementation of the Bed Management Application on Patient Satisfaction at Langsa General Hospital, integrating theoretical perspectives from the Technology Acceptance Model (TAM), the DeLone and McLean Information System Success Model (ISSM), and the SERVQUAL framework. Using a quantitative explanatory–predictive approach, the research employs both statistical regression analysis (SPSS 26.0) and algorithmic predictive modeling (Python Decision Tree Classifier) to measure and predict the relationship between system implementation and patient satisfaction. Data were collected from 120 inpatients who experienced the digital bed allocation process, using validated indicators that capture ease of use, reliability, accuracy, service speed, and transparency. The results of the regression analysis reveal that the implementation of the Bed Management Application has a positive and statistically significant effect on patient satisfaction (B = 0.687, β = 0.682, p < 0.001), with a coefficient of determination (R² = 0.465), indicating that 46.5% of the variance in satisfaction can be explained by system implementation effectiveness. Complementary algorithmic analysis using the Decision Tree Classifier achieved a prediction accuracy of 50%, identifying a key threshold at X_mean = 4.1, above which patients were predominantly classified into the High Satisfaction category. The findings confirm that technological quality, perceived usefulness, and information transparency significantly influence patient satisfaction, validating the theoretical constructs of TAM and ISSM. Furthermore, the integration of inferential and predictive analyses offers both theoretical validation and operational insight, illustrating that robust digital system implementation enhances patient experience, efficiency, and service reliability. This research contributes to advancing hybrid analytical approaches in health informatics, supporting data-driven decision-making and the national Smart Hospital Initiative to optimize patient-centered digital healthcare delivery in Indonesia.
The Best Caregiver at SOS Children’s Villages Using a Decision Support System Muhammad Iqbal; Syahputri, Maulisa
Journal of Computer Science and Research (JoCoSiR) Vol. 3 No. 1 (2025): Jan: Computer Science
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v3i1.67

Abstract

This study focuses on the development and implementation of a Decision Support System (DSS) designed to determine the best caregiver at SOS Children’s Villages. The main objective is to enhance efficiency and objectivity in the decision-making process related to caregiver performance evaluation. The research methodology includes collecting caregiver performance data, analyzing organizational needs, and applying an appropriate decision-making model. The DSS developed in this study utilizes Artificial Intelligence (AI) techniques to process and analyze performance data, generate performance scores, and identify the best caregiver based on predetermined criteria. The results show that the implementation of the DSS improves the objectivity of performance evaluations and provides significant support in the decision-making process. With this system, the organization is expected to better identify and optimize the potential of each caregiver, thereby increasing productivity and strengthening the competitiveness of SOS Children’s Villages in Medan. The collected data is processed and evaluated using the Simple Additive Weighting (SAW) method. The results are presented in the form of rankings and scores for each caregiver, facilitating a more accurate and transparent decision-making process. This study is expected to contribute positively to improving the efficiency and effectiveness of human resource management at SOS Children’s Villages.
Smarter School Labs: Fast and Accurate Anomaly Detection Using Lightweight CNN Technology Marbun, Ramlan; Iqbal, Muhammad
Journal of Computer Science and Research (JoCoSiR) Vol. 3 No. 1 (2025): Jan: Computer Science
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v3i1.70

Abstract

This study proposes a lightweight convolutional neural network (CNN) model for anomaly detection in school computer laboratories, aiming to enhance operational reliability and cybersecurity awareness. Real-time event logs were collected from 20 computers (PC01–PC20) at Santo Nicholas School with slight variations in CPU, RAM, and network behavior to simulate real-world heterogeneity. After preprocessing and normalization, the merged dataset contained over 10,000 log entries labeled as normal or anomalous. The proposed lightweight CNN achieved 92.23% F1-score, 91.80% accuracy, and a false positive rate (FPR) of 18.47%, demonstrating a balance between detection precision and computational efficiency. Comparative evaluation shows that this architecture performs competitively while requiring fewer parameters and lower inference latency than conventional CNNs. The results highlight the suitability of the proposed model for deployment in low-resource educational environments, supporting early anomaly detection and preventive maintenance. Future research will explore cross-domain generalization and lightweight deployment through edge-AI integration.
I-V Characterization and Electrical Performance Analysis of Undoped, N-Type, and P-Type Silicon for Semiconductor Applications Anyaora , Sunday Chimezie; Nwokeocha, Tochukwu Obialor; Nwaokafor, Innocent Chisom Chukwuma-; Takim , Stephen A.
Journal of Computer Science and Research (JoCoSiR) Vol. 2 No. 4 (2024): October: Artificial Intelligence
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65126/jocosir.v3i1.47

Abstract

Silicon remains the backbone of modern semiconductor technology; however, its intrinsic electrical limitations, such as low conductivity and restricted charge carrier concentration, constrain device performance. To enhance its functionality in electronic and photovoltaic applications, doping with suitable impurities is essential. This study focused on the I–V characterization and electrical performance analysis of undoped, N-type, and P-type silicon to assess the effect of doping on charge transport behavior. The experiment involved I–V characterization of intrinsic, N-type, and P-type silicon samples using precise materials, contact metals, and cleaning agents to ensure accuracy. A DC power supply, Source Measure Unit (Keithley 2400), and four-point probe station were employed for voltage application and current measurement. Samples were cleaned, coated with silver or aluminum contacts, annealed, and stored under nitrogen to prevent oxidation. I–V measurements were conducted under controlled environmental conditions, using calibrated equipment and multiple readings for accuracy. Data analysis in MATLAB included filtering, curve fitting, and extraction of key parameters like resistance and ideality factor to compare doped and undoped samples. The I–V characterization revealed clear differences between undoped and doped silicon samples. The undoped silicon exhibited Ohmic behavior with low conductivity ((5.3±0.2)×10⁻⁵ Ω⁻¹), while the N-doped ((1.4±0.1)×10⁻³ Ω⁻¹) and P-doped ((9.7±0.8)×10⁻⁴ Ω⁻¹) samples showed rectifying characteristics. N-type silicon displayed a lower turn-on voltage (0.65±0.02 V) than P-type (0.72±0.03 V), reflecting higher electron mobility. Ideality factors near unity (1.12 and 1.18) indicated diffusion-controlled transport. Conductivity improved 26-fold for N-type and 18-fold for P-type compared to intrinsic silicon, confirming doping’s strong influence on charge carrier concentration and validating measurement accuracy (standard deviation <3%). The study concludes that controlled doping significantly improves silicon’s electrical properties, making it more suitable for high-efficiency semiconductor and photovoltaic device applications.
Use of Data Warehouse and Data Mining for Academic Data: A Case Study at a National University Muhammad Iqbal; Muhammad Hasyim As’ary
Journal of Computer Science and Research (JoCoSiR) Vol. 3 No. 2 (2025): April: Artificial Intelligence
Publisher : Asosiasi Perguruan Tinggi Informatika dan Ilmu Komputer (APTIKOM) Provinsi Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Universities must optimize their information resources to enhance organizational performance and support strategic decision-making. However, academic data stored in multiple operational systems often remains fragmented and difficult to analyze comprehensively. This study aims to develop a data warehouse and apply data mining techniques to integrate and analyze academic data at the National University (UNAS), Jakarta. The data warehouse was designed using a star schema model, integrating academic records from various operational databases into a centralized repository. Mondrian and JPivot were utilized for multidimensional data presentation, while Classification-Based Association (CBA) and Association Rule techniques were applied to uncover hidden patterns within the data. The results show that the data warehouse significantly improves reporting efficiency, reducing processing time from one month to one day. Data mining analysis further revealed characteristic patterns among students in selecting specialization programs based on academic performance. These findings demonstrate that the integration of data warehousing and data mining supports more accurate reporting, informed decision-making, and data-driven academic planning.

Page 4 of 7 | Total Record : 63

 2   3   4   5   6 

Filter by Year

2023 2025


Reset
Filter By Issues
All Issue Vol. 3 No. 4 (2025): Research of Biotechnology Vol. 3 No. 3 (2025): July: Health Science Informatic Vol. 3 No. 2 (2025): April: Artificial Intelligence Vol. 3 No. 1 (2025): Jan: Computer Science Vol. 2 No. 4 (2024): October: Artificial Intelligence Vol. 2 No. 3 (2024): July: Artificial Intelligence Vol. 2 No. 2 (2024): April: Computer Science Vol. 2 No. 1 (2024): Jan: CNN and Artificial Vol. 1 No. 4 (2023): Oct: Computing Quantum and Related Fields Vol. 1 No. 3 (2023): July: Computer Science and Research Vol. 1 No. 2 (2023): April: Article Computer Science and Research Vol. 1 No. 1 (2023): January: Article Computer Science and Research More Issue