JISKA (Jurnal Informatika Sunan Kalijaga)

Peramalan Nilai Saham BBCA Melalui Pendekatan Time Series Menggunakan Teknik Exponential Smoothing

2024-01-31T14:18:01+07:00

Forecasting stock prices plays a crucial role in shaping investment strategies within the financial market. This article aims to predict the stock prices of Bank Central Asia (BBCA), a prominent entity in the Indonesian banking sector. Employing a time series methodology, this study utilizes the Exponential Smoothing technique to anticipate the fluctuations in BBCA's share prices. Meanwhile, the dataset used is the BBCA share price data from April 2001 to early January 2023. The final error rate in this forecast is 10%.

Enhancing Diabetes Classification Using a Relaxed Online Maximum Margin Algorithm

2025-02-23T19:43:20+07:00

Diabetes mellitus is a growing global health concern that requires accurate and reliable classification models for early diagnosis and effective management. Traditional machine learning models often struggle with class imbalance, generalization limitations, and high false-positive rates, leading to misdiagnoses and delayed interventions. This study enhances the Relaxed Online Maximum Margin Algorithm (ROMMA) to improve the accuracy of diabetes classification. Using a publicly available dataset from Kaggle, which contains 768 medical records with nine health attributes, the model’s performance was evaluated through a confusion matrix and classification metrics. The Enhanced ROMMA achieved an accuracy of 92%, significantly improving upon the Standard ROMMA’s 85% accuracy. The recall for diabetes detection increased from 0.83 to 0.94, reducing false negatives and ensuring more accurate patient identification. While slight misclassification still exists, this improvement enhances the model’s reliability for clinical applications. Future research should incorporate larger datasets and advanced techniques to enhance robustness and generalizability. This study contributes to the development of more accurate machine learning models for diabetes prediction, ultimately supporting better healthcare decision-making.

Analisis Ketertarikan Pengguna Microsoft Excel Online untuk Pengolahan Data Silsilah Keluarga Menggunakan TAM dan TPB

2024-10-23T17:21:36+07:00

The use of web-based applications such as Microsoft Excel Online has increased, including for recording family genealogy data. This study aims to analyze the factors influencing the intention and behavior of using this application based on the Technology Acceptance Model (TAM), Theory of Planned Behavior (TPB), and their combined framework. The constructs examined include perceived ease of use, perceived usefulness, attitude, subjective norm, perceived behavioral control, intention, and behavior. This quantitative study collected primary data through questionnaires distributed to family members using Microsoft Excel Online. Data analysis was conducted using SEM-PLS (Structural Equation Modeling-Partial Least Squares) with the assistance of SmartPLS version 4.1.0.2. The results indicate that perceived ease of use and perceived usefulness positively and significantly affect attitude, while attitude, subjective norm, and perceived behavioral control positively influence behavioral intention. Furthermore, behavioral intention has a positive effect on actual usage behavior. These findings suggest that Microsoft Excel Online is reliable for recording family genealogy data and supports technology acceptance among users.

Algoritma K-Means dan Analisis Komponen Utama untuk Mengatasi Multikolinearitas pada Pengelompokan Kabupaten Tertinggal

2024-08-09T11:38:13+07:00

Underdeveloped areas are regions that frequently face developmental challenges in various aspects such as infrastructure, education, and healthcare. Presidential Regulation Number 63 of 2020 designates 62 regencies in Indonesia as underdeveloped areas. This study categorizes the 62 underdeveloped regencies based on education and health indicators. The methods used are the k-means algorithm and principal component analysis due to multicollinearity in the data. MANOVA is conducted to determine the influence of the cluster results on the Human Development Index (HDI), Average Years of Schooling (AYS), Expected Years of Schooling (EYS), and Life Expectancy (LE). Due to multicollinearity in the education indicator data, principal component analysis was performed, resulting in three main components. The k-means analysis groups the 62 regencies into three clusters based on education indicators and two clusters based on health indicators. Further analysis using MANOVA shows the influence of the education and health clusters on HDI, AYS, EYS, and LE, indicated by statistical test results showing p-value < a(0.05). Thus, education and health indicators influence the categorization of underdeveloped areas.

Spatial Decision Support System to Determine the Feasibility of Evacuation Posts in Natural Disasters

2024-12-07T20:51:10+07:00

This study aimed to improve the accuracy of determining the feasibility of evacuation posts after natural disasters using the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) within a Spatial Decision Support System (SDSS). A dataset of 50 evacuation posts from the 2021 Mount Semeru eruption was analyzed. The Rank Order Centroid (ROC) method was applied for criteria weighting, and TOPSIS was used to process the data. Results showed 72% accuracy, confirming that TOPSIS is a passable method for assessing post-feasibility based on accessibility, sanitation, and refugee facilities. Although the focus is on evaluating post-disaster evacuation posts, the system can be adapted for use in various other types of disasters. However, it is still dependent on historical data and lacks real-time adaptability. Future research can integrate Artificial Intelligence (AI) and Machine Learning (ML) with real-time data to improve decision-making in disaster management.

Implementasi Metode YOLOv8 Mendeteksi Komputer Aktif dengan Subjek Layar Monitor

2025-02-07T23:48:17+07:00

Computers are one example of technological advances used in education. The use of computers that are not turned off can cause damage to computer components, and the use of electrical energy can increase. Student disobedience in turning off school laboratory computers when finished using them causes teachers to conduct manual checks by visiting each computer laboratory in the school. Deep learning is a machine learning algorithm that uses artificial neural networks. Deep learning is usually used for image recognition, voice identification, and data pattern analysis. Therefore, this study will apply the Deep Learning method, specifically YOLOv8, which aims to detect active computers based on the subject of the monitor screen and is expected to provide information about computers that are still active in the school laboratory. Based on the study's results, which detected 10 active computers, the 200-epoch model was selected with 100% accuracy at a speed of 2ms. Twenty active computers were selected, with 200 epoch models achieving 95% accuracy at a speed of 6ms per epoch. Thirty active computers were selected, with 100 epoch models achieving 96.67% accuracy at a speed of 3ms.

Klasifikasi Hewan Anjing, Kucing, dan Harimau Menggunakan Metode Convolutional Neural Network (CNN)

2024-09-13T21:10:24+07:00

Animal classification is a complex challenge due to variations in shape, color, and patterns across species. Traditional methods, which rely on manual feature extraction, are often ineffective in handling such complexities. Therefore, this study employs Convolutional Neural Networks (CNNs) as a more accurate approach for automatic feature extraction and image classification. This research aims to develop an animal image classification model, specifically for dogs, cats, and tigers, utilizing CNNs. The dataset consists of 4,800 images obtained from Kaggle, which were divided into training, testing, and validation sets. The CNN model was built using TensorFlow/Keras, trained for 50 epochs, and evaluated using accuracy, precision, recall, F1-score, and a confusion matrix. The experimental results show that the model achieved an overall accuracy of 88%, with the highest performance in tiger classification (99% accuracy). However, distinguishing between dogs and cats remains a challenge, with an accuracy of 81% for both classes. The findings indicate that CNNs are effective in automatically classifying animal images, although challenges persist in differentiating visually similar species. This study lays the groundwork for further enhancements, such as refining the model architecture or utilizing data augmentation techniques to boost classification accuracy.

Arsitektur Microservice untuk Optimalisasi Aplikasi Eco-Maps dalam Mendukung Kampus Ramah Lingkungan

2025-02-07T23:49:19+07:00

The implementation of environmentally friendly campus concepts has become increasingly crucial in addressing global environmental challenges. Eco-Maps is an application designed to visualize and manage sustainability efforts on campus, including energy management, waste management, and sustainable transportation initiatives. To enhance efficiency and flexibility, this study discusses the application of a microservice architecture in Eco-Maps. This architecture supports faster and more efficient development, testing, and deployment, while enabling horizontal scalability to manage high complexity and large data volumes. By separating application functions into independent services, microservices facilitate maintenance and updates while minimizing the impact of failures in individual services. This study also reviews the integration of containerization technologies, such as Docker and Kubernetes, to support microservice implementation. Through these technologies, the application can be deployed quickly and consistently across various environments, from development to production. System testing was conducted using load testing and stress testing methods, as shown in Tables 3 and 4. The results demonstrate that the average response time across ten iterations was 745.9 ms, with an average CPU usage of 44.38%. These findings confirm that processing load directly affects CPU efficiency and overall system performance.

Analisis Efektivitas Metode Filtering dan Intersection dalam Analisis Data Permukaan Bangunan dengan QGIS

2024-11-21T17:27:03+07:00

This study evaluates the efficiency of two methods for processing geospatial building surface data, namely Filtering and Intersection, using a case study in Blitar Regency. The data for this research was obtained by comparing two sources: OpenStreetMap (OSM), which has a data completeness rate of 60%, and Google Open Building, with a data completeness rate of 90%. From these two sources, the data with the highest completeness, which is from Google Open Building, was selected for further analysis. The data processing was carried out using QGIS software, chosen for its capability to support various geospatial analysis methods. The comparison of the two methods was based on three main criteria: processing time, resource efficiency, and scalability. The results showed that the Filtering method outperforms in all these aspects. Filtering can complete processing in an average of 1.6 seconds, significantly faster than the Intersection method, which requires an average of 7 minutes and 50 seconds. In terms of resource efficiency, Filtering is also more economical, with an average CPU usage of 18.85% and memory usage of 121.4 MB, compared to the Intersection method’s 34.05% CPU usage and 236.4 MB of memory. Additionally, the Filtering method demonstrated better scalability, capable of handling larger datasets with fewer resources and less time. Therefore, the Filtering method is recommended for geospatial data processing that prioritizes speed, efficiency, and the ability to handle large and complex datasets.

Analisis Sistem Deteksi Citra untuk Optimalisasi Pengawasan Lalu Lintas Udara Menggunakan Algoritma YOLOv5

2024-11-21T17:37:23+07:00

This study aims to develop an image detection system capable of identifying manned and unmanned aircraft objects to support air traffic surveillance. The increasing flight activity, both from commercial aircraft and drones, requires a more optimal surveillance system to connect the airspace efficiently. In this study, a Convolutional Neural Network (CNN) model utilizing the You Only Look Once version 5 (YOLOv5) method is employed to detect and classify objects in real-time from aircraft images. The methodology employed includes collecting aerial image data, labeling the data, and training object detection models using YOLOv5. The dataset used consists of 2,520 images of manned aircraft (warplanes) and 5,422 images of unmanned aircraft (drones). The experimental results demonstrate that the YOLOv5 model achieves high detection accuracy for both manned and unmanned aircraft, with a relatively fast inference time, thereby supporting the development of an effective air traffic surveillance system. This system is expected to be an integral part of a more sophisticated and responsive air traffic surveillance solution.