Jurnal Online Informatika
Not a member yet
276 research outputs found
Sort by
Case Study in Network Security System Using Random Port Knocking Method on The Principles of Availability, Confidentiality and Integrity
Full text linkPreventing unidentified individuals from misusing their access to information is a major concern when it comes to data security. Network administrators are charged with working harder to be able to secure the computer network they manage. The utilization of right method is a challenge for network administrators to protect computer network from intruders. The RPK method is one of solution to overcome this problem. This research aims to implement RPK method on the principles of availability, confidentiality, and integrity which have not been explored by previous studies. The network system configuration stage involved installing Debian 9, NMAP, Hydra, RPK, cloud server, remote admin, and attacker. The network security system\u27s performance was tested, revealing a 99.97% availability rate and 100% confidentiality. The system\u27s integrity was assessed, with an average response time of 0.22 seconds and 100% blocking accuracy. The test results indicate that the system\u27s network security performance, using the RPK method, capable of protecting server attacks and effectively upholding security stability
File Integrity Monitoring as a Method for Detecting and Preventing Web Defacement Attacks
Full text linkThe cybersecurity landscape in Indonesia recorded an increase in cyberattacks in 2022. One of the types of attacks observed was web defacement attacks targeting government websites. In 2022, there were a total of 2,348 web defacement attacks in Indonesia, with the majority occurring in the governmental sector. In proactive efforts to monitor and prevent web defacement attacks, this study implemented the open-source tool Wazuh and activated the file integrity monitoring module to detect file changes in the system. Testing was conducted with two types of attacks: brute force attacks to gain system access and web defacement attacks involving script insertion to trigger alerts from the file integrity monitoring. The results of the testing show that the implementation of Wazuh and the file integrity monitoring module can real-time detect malicious activities and file additions, so that it can be used to mitigate cyberattacks
Performance Comparative Study of Machine Learning Classification Algorithms for Food Insecurity Experience by Households in West Java
Full text linkThis study aims to compare the classification performance of the random forest, gradient boosting, rotation forest, and extremely randomized tree methods in classifying the food insecurity experience scale in West Java. The dataset used in this research is based on the Socio-Economic Survey by Statistics Indonesia in 2020. The novelty of this research is comparing the performance of the four methods used, which all are the tree ensemble approaches. In addition, due to the imbalance class problem, the authors also applied three imbalance handling techniques in this study. The results show that the combination of the random-forest algorithm and the random-under sampling technique is the best classifier. This approach has a balanced accuracy value of 65.795%. The best classification method results show that the food insecurity experience scale in West Java can be identified by considering the factors of floor area (house size), the number of depositors, type of floor, health insurance ownership status, and internet access capabilities
Improving with Hybrid Feature Selection in Software Defect Prediction
Full text linkSoftware defect prediction (SDP) is used to identify defects in software modules that can be a challenge in software development. This research focuses on the problems that occur in Particle Swarm Optimization (PSO), such as the problem of noisy attributes, high-dimensional data, and premature convergence. So this research focuses on improving PSO performance by using feature selection methods with hybrid techniques to overcome these problems. The feature selection techniques used are Filter and Wrapper. The methods used are Chi-Square (CS), Correlation-Based Feature Selection (CFS), and Forward Selection (FS) because feature selection methods have been proven to overcome data dimensionality problems and eliminate noisy attributes. Feature selection is often used by some researchers to overcome these problems, because these methods have an important function in the process of reducing data dimensions and eliminating uncorrelated attributes that can cause noisy. Naive Bayes algorithm is used to support the process of determining the most optimal class. Performance evaluation will use AUC with an alpha value of 0.050. This hybrid feature selection technique brings significant improvement to PSO performance with a much lower AUC value of 0.00342. Comparison of the significance of AUC with other combinations shows the value of FS PSO of 0.02535, CFS FS PSO of 0.00180, and CS FS PSO of 0.01186. The method in this study contributes to improving PSO in the SDP domain by significantly increasing the AUC value. Therefore, this study highlights the potential of feature selection with hybrid techniques to improve PSO performance in SDP
The Effect of the Number of Nodes on Data Communication Performance in Nomad Clusters Using the Gossip Protocol
Full text linkThis research aims to understand the effect of the number of nodes on the performance of data communication in Nomad clusters using the gossip protocol. Through a series of tests, it can be concluded that data communication performance is greatly affected by the number of nodes in the cluster. Tests were conducted using two clusters, where one cluster consists of three nodes. The results show that when using a cluster with three nodes, no packet loss occurs in all data transmissions performed, indicating a reliable communication system. The average latency in one data communication cycle varied in each test, but generally remained within the acceptable range of below 100ms based on data communication quality of service parameters. CPU and disc usage remained relatively stable throughout the experiment. Although there were slight differences in throughput between clusters, the throughput generally remained above 100 Mbps, which is still in the good category according to the research parameters. These results show the importance of taking into account the number of nodes in the cluster in designing and managing data communication systems in a Nomad cluster environment with the gossip protocol
Prediction of Solar Radiation Data for Garlic Production in Magelang Regency Using Long Short-Term Memory
Full text linkGarlic importation in Indonesia is frequently carried out to meet the high domestic market demand. To reduce dependency on imports, the development of local garlic production is crucial. This study aims to determine the optimal solar radiation for garlic growth using the Long Short-Term Memory (LSTM) algorithm. This algorithm was selected due to its ability to analyze time-series data and predict long-term patterns. The LSTM model was trained with the Adam optimizer, using a configuration of 1000 epochs, a batch size of 6, and a dropout rate of 2.0 to prevent overfitting. The model evaluation results show an indicating good accuracy with a RMSE of 0.1020, a Mean Squared Error (MSE) of 0.0104, and a correlation coefficient of 0.740, although it still has limitations in capturing extreme data fluctuations. The study found that in Magelang Regency especially in the sub-districts of Windusari, Grabag, Ngablak, Pakis, Dukun, Kaliangkrik, and Kajoran have optimal solar radiation for garlic cultivation between March and May, with a radiation range of 380 W/mยฒ to 440 W/mยฒ. These findings provide valuable guidance for farmers in determining the optimal planting period, potentially enhancing local garlic production and reducing import dependency
Enhancing Remote Sensing Image Quality through Data Fusion and Synthetic Aperture Radar (SAR): A Comparative Analysis of CNN, Lightweight ConvNet, and VGG16 Models
Full text linkRemote sensing technology benefits many parties, especially for carrying out land surveillance with comprehensive coverage without needing to move the equipment close to photograph the area. However, this technology needs to improve: the image quality depends on natural conditions, so objects such as fog, clouds, and smoke can interfere with the image results. This study uses data fusion techniques to enhance the quality of remote-sensing images affected by natural conditions. The method involves using Synthetic Aperture Radar (SAR) to combine adjacent satellite images from different viewpoints, thereby improving image coverage. Three image classification models were evaluated to process the fused data: Convolutional Neural Network (CNN), Lightweight ConvNet, and Visual Geometry Group 16 (VGG16). The results indicate that all three models achieve similar accuracy and execution speed, namely 0.925, with VGG16 demonstrating a slight superiority over the others, namely 0.90
Modeling Face Detection Application Using Convolutional Neural Network and Face-API for Effective and Efficient Online Attendance Tracking
Full text linkThe pandemic of Covid-19 emergency has ended, but it gives us a new lifestyle every aspect of life and also in the education aspect has changed. At that moment as one of the ways to prevent pandemic infection, many governments give the policy to close the offline class and continue with online classes. The online class system encountered several problems and one of those problems was to track the studentsโ attendance to ensure all the students were attending the class. The teacher needed extra effort to track it because they needed to call the students one by one which is wasting time and sometimes would miss the presence of the students who attend the class. To make it effective efficient accurate and time-consuming when tracking attendance in online classes for teachers, we proposed the face detection model which combines face-api.js and CNN to detect and recognize the studentsโ faces to help teachers track attendance by just uploading the screenshot image of the online meeting application. We tested our model with accuracy and speed testing. With 3 images of every studentโs face as training data, our model was able to recognize the face with 100% accuracy in just 41,65 seconds which is faster than calling students one by one that need almost 3 to 5 minutes if there are many students. Future research can be done by focusing research on improving the model to detect the studentsโ faces with different brightness, contrast, and saturation because students may not have the same place and condition when joining an online meeting class
Detection of Drowsiness in Drivers Using Image Processing and Support Vector Machine (SVM) Classification
Full text linkAccidents can be caused by external factors on the road, vehicle conditions, or internal factors such as drowsiness. Drowsiness while driving poses risks to the driver and others. An early detection system is crucial to alert drivers to stop or rest if they show signs of drowsiness. Physical signs of drowsiness include a lethargic facial expression, frequent eye blinking, continuous yawning, or nodding off. A detection system utilizing image processing and machine learning can observe these signs by detecting facial landmarks and analyzing activities such as eye blinking, yawning, and head tilt. This study aims to classify the drowsiness condition based on these three factors. The classification process is conducted using machine learning with the Support Vector Machine (SVM) method to determine whether a person is drowsy or not. The dataset consists of the number of eye blinks, head tilts, and yawns. Conditions are classified into two classes, drowsy and not drowsy. In this study, the SVM classification method can predict drowsiness with an accuracy of up to 77% in the conducted tests
A Mathematical Modelling and Behaviour Simulation of a Smart Grid Cyber-Physical System
Full text linkThe significant contributions of information and communication technology (ICT) and other operational technologies (OTs) or cyber networks have had a tremendous impact on the real-time monitoring, management, and control of the power or energy system facilities. Thus, the integration of these technologies into the energy grid system created a smart, complex, and interdependent system. This system is established and referred to as a smart grid cyber physical power system (SGCPPS). The performances of cyber physical systems are achieved via computation and communication and are imperatively based on a real-time feedback mechanism. In reference to the energy system, monitoring and control of the grid systems is extremely essential in ensuring efficient power supply, quality, reliability, stability and resilience among other determinants. However, their interdependence and integrated nature exposes the grid to disturbances subsequently leading to faults in the grid. Hence, failure to know the grid conditions at a particular period subjugates it to complete system collapse. This paper focused on the development of a mathematical model for a smart gird cyber physical system. Additionally, simulations were performed to study the behaviour of the Smart grid cyber-physical power system (SGCPPS) with regards to monitoring and controlling the physical systems using MATLAB Simulink tool to facilitate system awareness