cover
Contact Name
I Gede Surya Rahayuda
Contact Email
igedesuryarahayuda@unud.ac.id
Phone
+6289672169911
Journal Mail Official
jnatia@unud.ac.id
Editorial Address
Sekretariat JNATIA Gedung FMIPA Lantai 1, Program Studi Informatika, Fakultas Matematika dan Ilmu Pengetahuan Alam, Universitas Udayana
Location
Kota denpasar,
Bali
INDONESIA
Jurnal Nasional Teknologi Informasi dan Aplikasinya
Published by Universitas Udayana
ISSN : 29863929     EISSN : 30321948     DOI : -
JNATIA (Jurnal Nasional Teknologi Informasi dan Aplikasinya) adalah jurnal yang berfokus pada teori, praktik, dan metodologi semua aspek teknologi di bidang ilmu komputer, informatika dan teknik, serta ide-ide produktif dan inovatif terkait teknologi baru dan teknologi informasi. Jurnal ini memuat makalah penelitian asli yang belum pernah diterbitkan. JNATIA (Jurnal Teknologi Informasi dan Aplikasinya) diterbitkan empat kali setahun (Februari, Mei, Agustus, November).
Articles 71 Documents
Klasifikasi Citra Jamur Menggunakan SVM dengan PCA Berbasis Ekstraksi Fitur Hibrida I Putu Andika Arsana Putra; I Gusti Agung Gede Arya Kadyanan
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p02

Abstract

The general public still faces significant difficulty in differentiating between poisonous and non-poisonous mushrooms due to their high visual similarity. This has led to numerous poisoning incidents due to consumption of poisonous mushrooms. Between 2010 and 2020, there were 76 reported cases of poisoning involving 550 victims, 9 of whom died. To address this issue, a classification model was developed to differentiate between poisonous and non-poisonous mushrooms using Support Vector Machine (SVM) and Principal Component Analysis (PCA) algorithms based on hybrid feature extraction. The dataset for this study was obtained from Kaggle. The model built using PCA saw an increase in the model training time to 3 minutes 32 seconds from the initial 16 minutes 4 seconds without using PCA. Hyperparameter tuning was performed to find the best combination of parameters, resulting in RBF kernel, C value of 10, and gamma set to scale. The model was evaluated using a confusion matrix to determine accuracy and class-specific metrics. The model performed well, achieving 85% accuracy on the test data.  
Perbandingan Metode Clustering K-Means, GMM, dan DBSCAN Berbasis Fitur RFM Putu Nadya Putri Astina; I Wayan Supriana; I Made Satria Bimantara
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p07

Abstract

In the digital banking era, understanding customer behavior has become essential for delivering relevant services and maintaining competitiveness. This study aims to develop and evaluate customer segmentation models by leveraging an extended RFM (Recency, Frequency, Monetary) model, incorporating both behavioral and demographic attributes. Preprocessing, feature engineering, handling outliers, and standardization were done on the data using a dataset of 100,000 bank transaction records from Kaggle. DBSCAN, Gaussian Mixture Model (GMM), and K-Means were the three clustering techniques that were employed and contrasted. The clustering performance was evaluated using the Silhouette Score, Calinski-Harabasz Index (CHI), and Davies-Bouldin Index (DBI). The output of DBSCAN was too noisy to be useful in the business world, despite having the highest scores on Silhouette: 0,667 and lowest score on DBI: 0,396. K-Means offered the most interpretable segmentation with five ideal clusters (Silhouette: 0,308; DBI: 0,957; CHI: 6191), identifying customer groups ranging from highly active to potentially inactive. The findings highlight the synergy between transactional features and clustering algorithms in generating actionable insights for banking strategy.
Klasifikasi Cuaca Menggunakan Algoritma Fuzzy Mamdani dan CART I Dewa Ayu Agung Rai Ratna Karang; Ida Ayu Gde Suwiprabayanti Putra
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p14

Abstract

Weather prediction plays a vital role in sectors such as agriculture, transportation, and disaster mitigation. Extreme weather conditions can lead to unpredictable deviations that may cause significant harm to society. This study aims to predict weather conditions using the Fuzzy Mamdani algorithm and CART (Classification and Regression Tree). A total of 1,461 daily historical weather records from Seattle, United States, were obtained from the Kaggle website “Weather Prediction.” The fuzzy system was applied to convert numerical weather parameters—precipitation, maximum temperature, minimum temperature, and wind speed—into representative scores based on expert-defined rules. These fuzzy scores were then used as additional features in training the CART model to enhance weather classification accuracy. The dataset was split into 80% training data (1,168 records) and 20% testing data (293 records). Evaluation results show that the integration of Fuzzy and CART achieved an accuracy of 81.23% on the testing set, with high precision, recall, and F1 score for dominant categories such as sun and rain. This study demonstrates that the combination of fuzzy logic and decision trees is effective for weather classification based on historical data.
Analisis Sentimen Pengguna TikTok Terhadap Progres Pembangunan IKN Menggunakan LSTM dan FastText Gusti Agus Sakah Aditia; Luh Arida Ayu Rahning Putri
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p21

Abstract

The development of the Capital City of the Archipelago (IKN) is one of the national strategic projects that has generated various reactions and public opinions. Social media, especially TikTok, has become an important platform for people to voice their views through video content and interactive comments. This research aims to analyze the sentiment of TikTok users towards the progress of IKN construction using a deep learning approach, namely the Long Short-Term Memory (LSTM) algorithm. Sentiment data was collected from Kaggle as many as 1472 Indonesian comments, then processed through the stages of normalization, tokenization, and conversion into word embeddings. The LSTM model was designed and trained to classify sentiment into positive, negative, and neutral. The results of the analysis are expected to provide a comprehensive picture of public perception towards IKN, identify critical issues that are often discussed, and measure the level of public acceptance or rejection of this project. This research is expected to contribute to a better understanding of the dynamics of public opinion in the digital era.
Klasifikasi Berita pada Sektor Ekonomi Indonesia dari Bisnis.com Menggunakan SVM dan TF-IDF I Ketut Tangkas Agus Sucita; Made Agung Raharja
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p22

Abstract

The growth of digital news in Indonesia has seen a significant surge, especially in the economic sector. However, the abundance of information makes it difficult for users to find relevant news. This study aims to classify economic news using the Support Vector Machine (SVM) method and Term Frequency-Inverse Document Frequency (TF-IDF). The process begins with collecting economic news data, preprocessing the text, feature extraction using TF-IDF, and classification using SVM.  Evaluation results show that the model achieved a training accuracy of 81.33% , a test accuracy of 76.57% , and an average 5-fold cross-validation accuracy of 77.3%. This indicates high accuracy and stable generalization in classifying news into appropriate categories. This study is expected to assist readers and information systems in efficiently filtering news content.
Optimasi Algoritma KNN Menggunakan Metode PSO dalam Klasifikasi Kanker Paru-Paru Made Arief Budi Dharma; I Made Widiartha
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p08

Abstract

Lung cancer remains one of the leading causes of cancer-related mortality worldwide, with early detection being critical for improving patient outcomes. However, conventional diagnostic methods often require substantial time and resources. This study implements and evaluates the integration of Particle Swarm Optimization (PSO) with the K-Nearest Neighbors (KNN) algorithm for lung cancer risk classification using a lung cancer dataset consisting of 20,000 samples and 16 predictive features. The study addresses KNN's limitation in handling irrelevant or redundant features, which can reduce classification accuracy. PSO, a population-based optimization algorithm inspired by the social behavior of bird flocks, is employed to perform feature selection, identifying the most relevant subset of features to enhance model performance. The results show that PSO successfully reduces the number of features from 16 to 4 with improve accuracy of 87,61%, over the baseline KNN model. This reduction improves computational efficiency and facilitates model interpretability without compromising performance, supporting the application of KNN-PSO as a decision support system for early lung cancer detection in clinical settings.
Implementasi Model Poisson Laplace untuk IR pada E-Skripsi Universitas Udayana I Putu Andhika Ardianta Putra; Cokorda Pramartha
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p13

Abstract

At Udayana University, a digital repository system is available to store books and student theses. However, the current repository system only categorizes documents based on the year of publication, without any grouping based on topics, fields of science, or other data. Although a search feature is available, the function has not been optimized. To overcome this problem, research was conducted on the Probabilistic Information Retrieval Model, namely the Poisson Model with Laplace Smoothing and Normalization 2. Research was conducted on student thesis title data in 2024 as many as 4,671 titles and evaluation using 10 queries. The research resulted in a Mean Average Precision value, normalized Discounted Cumulative Gain, and recall of 0.715 and a precision value of 0.1421. From this value, further experiments need to be held because the precision value is far different from the recall, indicating the number of False Positive values.
Evaluasi ICA dan NMF pada Pemisahan Sinyal Audio Menggunakan BSS Metrics dan MFCC Ni Ketut Sukardiasih; I Gede Arta Wibawa
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p16

Abstract

Source separation is a crucial challenge in audio signal processing, particularly for stereo data. This study compares the performance of Independent Component Analysis (ICA) and Non-negative Matrix Factorization (NMF) in separating mixed audio signals. ICA operates directly on stereo signals, while NMF is applied to mono versions derived from stereo mixtures. Three pairs of audio data with diverse natural sound combinations were used. Evaluation metrics include Blind Source Separation indicators (SDR, SIR, SAR), spectral similarity based on Mel-Frequency Cepstral Coefficients (MFCC), and robustness tests by adding noise at 10 dB and 5 dB SNR levels. The results show that ICA consistently yields higher SDR and SIR scores and lower Euclidean distances in MFCC compared to NMF. In contrast, NMF performs poorly due to its mono-only limitation and inability to exploit spatial information. This study highlights ICA's superiority in separation accuracy and noise robustness, and emphasizes the importance of spectral analysis as a complementary evaluation method.
Perancangan Sistem Notifikasi Tamu Menggunakan Kombinasi Sensor PIR dan Ultrasonik Dewa Putu Adrian Paramarta; I Made Widhi Wirawan
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p09

Abstract

Modern lifestyles often cause homeowners to be unaware of a guest's presence due to distracting activities like working from home, gaming, or using social media. Conventional doorbells fall short in these situations, highlighting the need for a more effective, technology-based solution. This research focuses on designing and building an innovative guest notification system based on the Internet of Things (IoT) using an ESP32 microcontroller. To enhance accuracy and reduce false positives, the system integrates a dual-sensor verification mechanism. A Passive Infrared (PIR) sensor acts as the initial trigger to detect motion, which is then verified by an HC-SR04 ultrasonic sensor to confirm the object's presence within a relevant distance from the door. A notification is sent via the Telegram application only when both conditions are met. This approach effectively filters out irrelevant movements and significantly improves reliability compared to single-sensor systems. The final prototype demonstrates an accurate, affordable, and practical solution to ensure homeowners never miss a visitor.
Perbandingan CNN dan SVM untuk Klasifikasi Citra Rempah Indonesia Komang Arjuntama Satria; I Wayan Supriana
Jurnal Nasional Teknologi Informasi dan Aplikasinya Vol. 4 No. 2 (2026): JNATIA Vol. 4, No. 2, Februari 2026
Publisher : Informatics Department, Faculty of Mathematics and Natural Sciences, Udayana University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24843/JNATIA.2026.v04.i02.p17

Abstract

The purpose of this study is to evaluate how well Convolutional Neural Networks (CNN) and Support Vector Machines (SVM) perform while using digital image data to classify different kinds of Indonesian spices. This study's dataset, which includes a variety of spice images with different shapes, textures, and colors, was sourced from the Kaggle platform. CNN is frequently used in difficult picture classification problems and is well known for its capacity to automatically extract visual information. SVM, on the other hand, is a traditional machine learning method that has demonstrated reliable results in a range of classification tasks, especially when dealing with sparse data and well-organized features. There are several processes in the study technique, such as gathering data, preprocessing images, training the model, evaluating the model, and comparing performance. The models abilities and flaws in categorizing spice photos are thoroughly examined by evaluating them using important performance measures such as accuracy, precision, recall, and F1-score. The findings are anticipated to aid in the development of intelligent agricultural systems, particularly in automating the process of classifying and identifying spice items. Furthermore, this comparison can be used as a reference to choose suitable machine learning techniques for comparable picture classification problems including datasets linked to agriculture or food.