cover
Contact Name
Sarida Sirait
Contact Email
saridasrt@gmail.com
Phone
+6281319494217
Journal Mail Official
saridasrt@gmail.com
Editorial Address
Jl. Sriwijya No. 9 C-E Pematangsiantar, Sumatera Utara
Location
Kota pematangsiantar,
Sumatera utara
INDONESIA
Jurnal Tekinkom (Teknik Informasi dan Komputer)
ISSN : 26211556     EISSN : 26213079     DOI : https://doi.org/10.37600/tekinkom
Core Subject : Science,
Jurnal TEKINKOM merupakan jurnal yang dimaksudkan sebagai media terbitan kajian ilmiah hasil penelitian, pemikiran dan kajian analisis-kritis mengenai isu Ilmu - ilmu komputer dan sistem informasi, seperti : Pemrograman Jaringan, Jaringan Komputer, Teknik Komputer, Ilmu Komputer/Informatika, Sistem Informasi, dan Multi Disiplin Penunjang Domain Penelitian Komputasi, Sistem dan Teknologi Informasi dan Komunikasi, dan lain-lain yang terkait. Artikel ilmiah dimaksud berupa kajian teori (theoritical review) dan kajian empiris dari ilmu terkait, yang dapat dipertanggungjawabkan serta disebarluaskan secara nasional maupun internasional.
Articles 407 Documents
ANALISIS PERBANDINGAN MODEL RANDOM FOREST DAN GRADIENT BOOSTING MACHINE DALAM AKURASI PREDIKSI FAKTOR RISIKO PENYAKIT JANTUNG Sinaga, Michael Kevin; Aisyah, Siti; Sitanggang, Elsa Pricila; Sarumaha, Reimun; Amalia, Amalia; Radhi, Muhammad
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.1980

Abstract

mortality rates. This study aims to analyze and compare the performance of two machine learning algorithms—Random Forest and Gradient Boosting Machine (GBM)—in predicting heart disease risk based on patient medical data. A quantitative approach was used, incorporating Exploratory Data Analysis (EDA), data preprocessing, modeling, and evaluation using metrics such as accuracy, precision, recall, and F1-score. The dataset was obtained from Kaggle and included clinical attributes such as age, gender, blood pressure, cholesterol level, and chest pain type. The results show that both algorithms achieved high classification performance, with GBM outperforming overall, achieving 98.3% accuracy, 97.4% precision, 99.4% recall, and 98.4% F1-score. Meanwhile, Random Forest demonstrated strong performance with an accuracy of 94.7%. The most influential features in prediction were ST slope, oldpeak, and chest pain type. This study concludes that the application of GBM is more effective in supporting early heart disease detection and can serve as a fast, accurate, and efficient decision support system in healthcare settings with limited computational resources.
ANALISIS DAMPAK STRATEGI PEDAGOGI TERHADAP MINAT BELAJAR SISWA MENGGUNAKAN RANDOM FOREST Sinaga, Novendra Adisaputra; Pardede, Doughlas; Riyadi, Sugeng
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.2169

Abstract

Student learning interest plays a crucial role in educational success, as it directly influences engagement, comprehension, and academic achievement. This study aims to analyze the influence of pedagogical strategies on students’ learning interest using a machine learning approach with the Random Forest algorithm. Eight aspects of teaching strategies were examined as predictor variables, while learning interest was measured through two main indicators: interest in real-world application of the material and motivation for self-directed learning. Data were collected from 100 students via a Likert-scale questionnaire and analyzed using Orange Data Mining. The model was validated through 10-fold cross-validation and evaluated using accuracy, precision, recall, F1-score, and AUC. The results indicate strong model performance, with 95% accuracy, 96.7% precision, 97.8% recall, and a 97.2% F1-score. Feature importance analysis identified practical activities (P4), an inclusive learning environment (P6), and the use of technology (P3) as the most influential predictors of learning interest. In contrast, variables such as P1, P2, and P8 showed minimal contribution. These findings demonstrate that Random Forest is not only effective for classification tasks but also valuable in identifying key factors for improving pedagogical strategies. The results are expected to inform the development of more adaptive, interactive, and student-centered learning environments.
PREDIKSI METODE PERSALINAN DENGAN BIG DATA DAN ALGORITMA GRADIENT BOOSTING CLASSIFIER Fitriyani, Intan Nur; Fadillah, Riszki; Adawiyah, Quratih; D, Novica Jolyarni
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.1557

Abstract

This study aims to develop a prediction model to determine the method of delivery (normal or cesarean) using the Gradient Boosting algorithm based on maternal examination data. This model was evaluated using precision, recall, F1-score, and accuracy metrics. The results showed that the Gradient Boosting model had an accuracy of 48%, with better performance in predicting Normal delivery compared to Caesarean. Although this model is effective, there is an imbalance in precision and recall for the Caesarean class, indicating the need for improvement in identifying cases of cesarean delivery. Comparison with other algorithms such as Random Forest, Logistic Regression, and SVM showed that Random Forest gave the best performance with an accuracy of 55%. To improve performance, this study recommends hyperparameter optimization, application of class balancing techniques, and enrichment of medical features. The developed model has the potential to be used as a tool in medical decision-making related to delivery methods, which is expected to improve the safety of mothers and babies, and reduce dependence on subjective factors in medical decisions.
ANALISIS PROSES BISNIS DAN PERANCANGAN SISTEM INFORMASI KEUANGAN PADA RESTAURANT INDIAN FOOD DELHI6 BERBASIS MICROSOFT EXCEL VB Paramitha, I Gusti Agung Diah Pradnya; Juliharta, I Gede Putu Krisna; Alam, Helmy Syakh
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.1890

Abstract

Accounting information systems are the backbone of every company, including restaurants. Restaurant Indian Food Delhi6, as a fastfood restaurant operating in Bali, currently still relies on manual methods to record and manage its financial transactions. Although this manual method was effective in the past, along with the increasing number of transactions and business complexity, it began to show a number of weaknesses. To overcome these problems, this study aims to design and develop an integrated VBA Excel-based accounting information system specifically for Restaurant Indian Food Delhi6. This system is expected to automate various repetitive accounting tasks, such as recording sales, purchases, and payroll, thereby reducing employee workload and minimizing human error. In developing this system, the Waterfall and UML methods will be used. The Waterfall method will be used as a framework for overall system development, while UML will be used to model and document various aspects of the system, such as use cases, flowcharts, and class diagrams. The final result of this study is a comprehensive accounting information system design, including use cases, narratives, and an integrated financial system. This system is expected to meet the specific needs of Restaurant Indian Food Delhi6 and be an effective solution to overcome current problems.
PERBANDINGAN ALGORITMA NAIVE BAYES & K-NEAREST NEIGHBORS (KNN) DALAM ANALISIS SENTIMEN ULASAN PRODUK TOKOPEDIA Purba, Windania; Turnip, Charles Fransisco; Malau, Josua Heksa Parti; Halawa, Berkat Editar Jaya
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.1983

Abstract

This study conducts a comparative performance analysis of two widely utilized classification algorithms, Naive Bayes and K-Nearest Neighbors (KNN), in the context of customer satisfaction analysis based on product reviews from the Tokopedia e-commerce platform. Customer-generated reviews serve as a critical factor in shaping product reputation and perceived quality, while also influencing the purchasing behavior of prospective buyers.The methodology encompasses data collection of product reviews from Tokopedia, followed by a comprehensive preprocessing pipeline, including text cleaning, tokenization, and stemming. The processed reviews are then categorized into two sentiment classes-positive and negative-employing both Naive Bayes and KNN algorithms.The performance of these algorithms is evaluated using standard classification metrics: accuracy,recall,F1-score dan precision. Empirical results demonstrate that Naive Bayes yields superior accuracy in classifying product sentiments compared to KNN.This research offers practical insights for e-commerce businesses in selecting suitable machine learning techniques for sentiment analysis to better understand customer feedback and enhance satisfaction. Moreover, the study contributes to the academic discourse by highlighting the strengths and limitations of each algorithm, and provides recommendations for future research in developing effective sentiment classification frameworks for customer satisfaction measurement.
PREDIKSI RISIKO KEBAKARAN MENGGUNAKAN ALGORITMA NAÏVE BAYES BERDASARKAN DATA HISTORIS DAN LINGKUNGAN Ismanizan, Ryan; Fitri, Triyani Arita; Rahmiati, Rahmiati; Agustin, Agustin
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.2352

Abstract

Fire is a disaster that can cause significant material and human losses. Kampar Regency in Indonesia is a fire-prone area due to short circuits, human negligence, and environmental conditions. This study aims to predict fire risk based on historical fire incident data and environmental factors using the Naïve Bayes algorithm. This method was chosen because of its ability to classify large-dimensional data with high probability. The research stages include data collection, preprocessing, data exploration, modeling, and model evaluation. Data were tested using splits of 70:30, 80:20, and 90:10. The results showed that the Naïve Bayes algorithm was able to provide prediction accuracy levels of 95.82%, 96.00%, and 95.45%, respectively. The highest accuracy level was obtained in the 80:20 scenario. These findings indicate that Naïve Bayes is effective in predicting high-risk areas for fire and can serve as a reference for relevant parties in developing more targeted fire prevention and mitigation policies.
OPTIMALISASI LAYANAN BUS TRANS METRO DELI MEDAN MELALUI CLUSTERING DATA MINING DENGAN METODE PARTITIONING AROUND MEDOIDS Sinambela, Rio; Matondang, Matius; Rizky, Rahmad; Tamba, Saut Parsaoran
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.1974

Abstract

Bus transportation in Indonesia plays a vital role in reducing traffic congestion and supporting sustainable urban mobility. This study aims to optimize the services of Trans Metro Deli Medan by analyzing operational patterns using the Partitioning Around Medoids (PAM) clustering method. The research applies a data mining approach to segment service performance based on passenger volume, travel time, departure frequency, and passenger satisfaction. The analysis results in three distinct clusters: Quiet (average 15 passengers, 38 minutes travel time, low frequency and satisfaction), Medium (50 passengers, 75 minutes, moderate frequency and satisfaction), and Dense (95 passengers, 105 minutes, high frequency and satisfaction). These clusters reveal strong correlations among operational variables and highlight disparities in service quality across routes. Based on the findings, service improvement strategies are proposed: enhancing service quality and visibility for quiet routes, maintaining operational stability for medium routes, and increasing fleet capacity with adjusted scheduling for dense routes. The study demonstrates that PAM clustering provides valuable insights for strategic planning and can significantly contribute to the efficiency and quality improvement of public bus services in Medan.
ANALISIS SENTIMEN TERHADAP ULASAN PENGGUNA APLIKASI DANA DI GOOGLE PLAYSTORE MENGGUNAKAN ALGORITMA SUPPORT VECTOR MACHINE (SVM) Willyarnandi, Muchammad Chadavi; Huizen, Lenny Margaretta
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.2253

Abstract

In today's era, technological advancements have significantly facilitated various aspects of life, including the financial sector through the emergence of financial technology (fintech). One of the widely used fintech services in Indonesia is the DANA digital wallet. The abundance of user reviews generated from the use of this application reflects the level of user satisfaction or dissatisfaction with the services provided. However, this data is generally unstructured text, making it difficult to analyze manually. Therefore, an automatic analysis method is needed to categorize the sentiments contained in these reviews. Support Vector Machine (SVM) is one of the algorithms that can be used for sentiment classification, although its effectiveness in analyzing reviews of the DANA application still requires further investigation.This study aims to analyze sentiment in user reviews of the DANA application obtained from the Google Play Store using the SVM algorithm. A total of 2,000 reviews were collected through scraping and processed through several stages, including data cleaning, case folding, normalization, tokenization, stopword removal, and stemming. Text features were extracted using the Term Frequency-Inverse Document Frequency (TF-IDF) method before classification with SVM. SVM was chosen due to its ability to handle high-dimensional data and its strong performance in text classification tasks. The results indicate that SVM is capable of classifying sentiment with high accuracy, achieving around 90% or more, along with high precision, recall, and F1-score values. These findings are expected to help application developers understand user needs and complaints to enhance the quality of DANA’s services.
SISTEM PENDUKUNG KEPUTUSAN PEMILIHAN MAHASISWA PKL TERBAIK MENGGUNAKAN METODE MOOSRA Purba, Arifin Tua; Manalu, Andi Setiadi; Sirait, Erwin
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.2137

Abstract

Politeknik Bisnis Indonesia is a vocational higher education institution committed to producing graduates who are not only academically competent but also equipped with practical skills required in the workforce. One of the essential programs in its curriculum is the Internship (PKL), designed to allow students to apply their knowledge in real-world work environments. However, the selection of the best internship students has been conducted manually, leading to inefficiencies and potential subjectivity in the evaluation process. This study aims to design a Decision Support System (DSS) using the MOOSRA (Multi-Objective Optimization on the Basis of Simple Ratio Analysis) method to support a more objective and systematic selection process. The evaluation involves five main criteria: discipline, teamwork, skill, work quality, and attendance, with six student candidates as alternatives. The research stages include problem identification, criteria and weight determination, data collection, data processing with the MOOSRA method, system design, and system testing. The results show that the MOOSRA method effectively ranks the students, with student A4 selected as the best internship participant with the highest Yi score of 6.12347. This research demonstrates that the MOOSRA method can significantly improve decision-making accuracy and fairness in multi-criteria selection processes.
PERBANDINGAN ALGORITMA RANDOM FOREST DAN SUPPORT VECTOR MACHINES DALAM MEMPREDIKSI TINGKAT RISIKO SERANGAN JANTUNG BERDASARKAN KEBIASAAN MEROKOK Harmaja, Okta Jaya; Fernando, Fernando; Melati, Melati
Jurnal Teknik Informasi dan Komputer (Tekinkom) Vol 8 No 1 (2025)
Publisher : Politeknik Bisnis Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37600/tekinkom.v8i1.1807

Abstract

Heart disease remains a major global health challenge, with smoking behavior identified as one of the most significant modifiable risk factors. This study aims to compare the performance of two machine learning algorithms—Random Forest and Support Vector Machine (SVM)—in predicting heart attack risk levels based on smoking habits and biometric indicators. Using a dataset of 3,901 subjects obtained from Kaggle, data preprocessing and feature engineering were conducted to optimize model accuracy. The SVM algorithm achieved an accuracy of 92.43%, with its best performance observed in the medium-risk category (precision: 0.95, recall: 0.97, F1-score: 0.96), although performance declined in low and high-risk categories. In contrast, the Random Forest algorithm demonstrated superior results, reaching 99.91% accuracy with perfect precision, recall, and F1-scores (1.00) across all risk categories. The findings indicate that Random Forest not only provides more consistent and accurate predictions but also minimizes classification errors effectively. This research suggests that Random Forest is a more reliable and robust approach than SVM for integrating into intelligent health information systems to support early detection and prevention strategies for heart disease, especially among individuals with active smoking behavior.