Claim Missing Document
Check
Articles

Found 35 Documents
Search

Comparison of Distance Measurements Based on k-Numbers and Its Influence to Clustering Deny Jollyta; Prihandoko Prihandoko; Dadang Priyanto; Alyauma Hajjah; Yulvia Nora Marlim
MATRIK : Jurnal Manajemen, Teknik Informatika dan Rekayasa Komputer Vol. 23 No. 1 (2023)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/matrik.v23i1.3078

Abstract

Heuristic data requires appropriate clustering methods to avoid casting doubt on the information generated by the grouping process. Determining an optimal cluster choice from the results of grouping is still challenging. This study aimed to analyze the four numerical measurement formulas in light of the data patterns from categorical that are now accessible to give users of heuristic data recommendations for how to derive knowledge or information from the best clusters. The method used was clustering with four measurements: Euclidean, Canberra, Manhattan, and Dynamic Time Warping and Elbow approach for optimizing. The Elbow with Sum Square Error (SSE) is employed to calculate the optimal cluster. The number of test clusters ranges from k = 2 to k = 10. Student data from social media was used in testing to help students achieve higher GPAs. 300 completed questionnaires that were circulated and used to collect the data. The result of this study showed that the Manhattan Distance is the best numerical measurement with the largest SSE of 45.359 and optimal clustering at k = 5. The optimal cluster Manhattan generated was made up of students with GPAs above 3.00 and websites/ vlogs used as learning tools by the mathematics and computer department. Each cluster’s ability to create information can be impacted by the proximity of qualities caused by variations in the number of clusters.
Analisis Penerapan Augmented Reality Sebagai Strategi Pemasaran: Uji Black Box dan Korelasi Kody, Jeffry; Jollyta, Deny; Hajjah, Alyauma; Pratama, Teddy
The Indonesian Journal of Computer Science Vol. 11 No. 1 (2022): The Indonesian Journal of Computer Science
Publisher : AI Society & STMIK Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33022/ijcs.v11i1.3037

Abstract

Traditional marketing no longer ensures greater revenue. Businesses in the advertising industry are also feeling the effects of this circumstance. People with a lot of mobility have less time to go shopping and visit stores. The demand for seeing product designs continues to rise, yet many people are unable to attend in person, resulting in greater time spent at work. Entrepreneurs must alter their marketing strategy to address these issues utilizing technology that is simple to use and available at all times. The goal of this research is to design an Augmented Reality (AR) application that can be used on a smartphone and can process sales via the internet. Black Box, light intensity, and the proper distance are used to create and test applications for functioning so that product photos seem at their best. The existence of the app also generates a strong correlation between customer interest of product and desire to purchase it. This is demonstrated by a correlation test with a value of 0.673191. It is envisaged that the designed application can aid advertising enterprises in enhancing marketing and sales.
Naïve Bayes and K-Nearest Neighbor Algorithm Approach in Data Mining Classification of Drugs Addictive Diseases Priyanto, Dadang; Iman, Ahmad Robbiul; Jollyta, Deny
ILKOM Jurnal Ilmiah Vol 15, No 2 (2023)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v15i2.1544.262-270

Abstract

Indonesia, with its very large population, is a potential market for drugs trafficking. Hence, seriousness is needed in cracking down or preventing drug trafficking. Narcotics are substances or drugs that can cause dependence or addicted and other negative impacts on users. The problem is that drug users do not realize and even ignore diseases caused by drug addiction. The diseases can be life-threatening for users, such as inflammation of the liver, heart disease, hypertension, stroke, and others. The prevalence rate of drug abuse in West Nusa Tenggara (NTB) is included in the high category, reaching 292 cases or around 37.24% cases. This study aimed to create an application that can classify various diseases of drug users using the naïve bayes and KNN methods. The results of this study indicated that there was a very close relationship between drug users and various deadly diseases. The prediction results showed that the naive bayes method provided a prediction accuracy of 94.5% while the KNN showed a prediction accuracy of 92.5%. This shows that the naive bayes method provides better predictive performance than the KNN in the data set of drug addicts in NTB.
Machine Learning Decision Support System for Heart Disease Prediction with Optuna and Threshold Optimization Ramdhan, William; Hutahaean, Jeperson; Jollyta, Deny; Karim, Abdul
Jurnal Teknik Informatika (Jutif) Vol. 7 No. 2 (2026): JUTIF Volume 7, Number 2, April 2026
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2026.7.2.5684

Abstract

Cardiovascular disease remains a major global health challenge, necessitating accurate and reliable decision support systems for early detection. This study proposes a machine learning–based decision support system that integrates ensemble learning, automated hyperparameter optimization using Optuna, and decision threshold tuning. The system was evaluated using several baseline machine learning models, including Logistic Regression, SVM, KNN, Decision Tree, and Random Forest, with the Random Forest model selected for optimization. Hyperparameter tuning with Optuna and decision threshold optimization led to a significant improvement in accuracy (95.0%) and ROC–AUC (0.977), with the optimized model outperforming all baseline models. This approach demonstrates improved sensitivity, reduced false negatives, and enhanced predictive performance, offering a clinically reliable tool for early heart disease detection. The results emphasize the importance of model optimization and decision threshold calibration in clinical decision support systems.
Enhancement of Supervised Learning Models for Intrusion Detection Through Mutual Information and Hyperparameter Tuning Jollyta, Deny; Makaruku, Yoakhina Nicole; Hajjah, Alyauma; Marlim, Yulvia Nora
MATRIK : Jurnal Manajemen, Teknik Informatika dan Rekayasa Komputer Vol. 25 No. 2 (2026)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/matrik.v25i2.5760

Abstract

Enhancing the performance of supervised learning algorithms through feature and hyperparameter testing remains challenging for users, particularly when detecting computer network intrusions. There are opportunities to assess whether a supervised learning algorithm performs optimally, depending on the number of features and the choice of hyperparameters. The purpose of this research is to enhance the network intrusion detection performance of three supervised learning algorithms, namely Support Vector Machine (SVM), eXtreme Gradient Boosting, and Random Forest, by using the Mutual Information feature selection approach and hyperparameter tuning. Mutual Information measures the dependency of features on the target. Features with high values are the most informative. Hyperparameters are not learned from the data; they are set before training begins. Hyperparameters are selected in accordance with the requirements of the three algorithms via iterative training and testing on the NSL-KDD dataset. The dataset was split into 80:20, 70:30, and 60:40. The results showed that the fifteen features with the highest mutual information were identified and trained on the data using appropriate hyperparameters. By splitting the data in an 80:20 ratio, the accuracy of Support Vector Machine reached its maximum, increasing from 90% to 98%. In contrast, eXtreme Gradient Boosting and Random Forest reached their maximum, increasing from 97% and 98% to 100%, respectively. The study’s findings advance our understanding of how algorithm performance depends on feature and hyperparameter selection.