Sinkron : jurnal dan penelitian teknik informatika
Not a member yet
1261 research outputs found
Sort by
Usability Testing of Industrial Engineering UPNVJT Website Using Eye Tracking and System Usability Scale
No full textWebsites are essential for agencies and organizations to ensure their information is accessible to the public. The eye tracker method is effective for evaluating user usability. The Industrial Engineering Study Program at Universitas Pembangunan Nasional "Veteran" East Java has an official website, tekindustri.upnjatim.ac.id, which has not been tested for usability. Initial problem identification revealed issues such as inaccessible menus and hidden information. This study aims to assess and improve the user experience on the website using a combination of Eye Tracking and System Usability Scale (SUS) methods. The average effectiveness score for 39 respondents is 89.10%, with 8% rated as ineffective, 28% as effective enough, and 64% as very effective. The efficiency value, measured across 39 respondents and 4 tasks, is 0.0276 goals/second, indicating each respondent completes 2.76% of tasks per second, requiring about 36.23 seconds to reach 100% task completion. The initial SUS score was 69.49%. Five issues were identified in the Home, Facilities, Education, Thesis, and MBKM sections and one design issue. A prototype was developed and tested, resulting in a final SUS score of 80.06%, placing the website in the marginally high acceptability range, category B for grade scale, and excellent for adjective ratings. The SUS score improvement was 10.57%. This research shows that combining Eye Tracking and SUS is an effective method for increasing website usability. The implications of this research can help organizations improve the quality of their websites and provide a better user experience
Classification of Breast Cancer with Transfer Learning on Convolutional Neural Network Models
No full textBreast cancer is a serious medical condition and a leading cause of death among women. Early and accurate diagnosis is crucial for improving patient outcomes. This study explores the use of Convolutional Neural Networks (CNNs) with Transfer Learning using DenseNet121 and ResNet50 models to enhance breast cancer classification via mammography. Transfer Learning enables CNN models to leverage knowledge learned from larger datasets such as ImageNet to improve performance on specific breast cancer datasets. The dataset comprised medical images with three breast variations: benign, malignant, and normal, totaling 531 data points. Data was split with a 70% training and 30% validation ratio. Two CNN models, AlexNet and ResNet50, were evaluated to compare their performance in classifying these breast cancer types. The experimental results show that AlexNet achieved a training accuracy of 98.01%, while ResNet50 achieved 64.07%. AlexNet demonstrated superior performance in identifying complex patterns in mammography images, resulting in more accurate classification of different breast cancer types. These findings highlight the potential of deep learning applications to support more precise and effective medical diagnostics for breast cancer. This research contributes significantly to the development of AI technologies in healthcare aimed at improving early detection of breast cancer. The implications of this study could expand our understanding of Transfer Learning applications in medical contexts, driving further advancements in this field to enhance patient care and prognosi
Lung Cancer Classification Using Combination Of Efficientnet And Visual Geometry Group Algorithm
No full textLung cancer is one of the leading causes of mortality All around the world. It is classified into three main types: Adenocarcinoma of the lung (ACA), Non-small cell lung cancer (N), and Squamous Cell Carcinoma of the lung (SCC). Lung Cancer Classification is crucial on development of effective treatments. This study aims to improve the accuracy of lung cancer classification through the integration of a hybrid model, which combines two Convolutional Neural Networks architectures, namely EfficientNet-B7 and VGG-16. A set of histopathology images was subjected to testing, with the data split into three categories: 60% for training, 30% for validation, and 10% for testing. Prior to use, each image underwent a preprocessing process, wherein it was resized to 256x256 pixels. The model test results achieved an accuracy, precision, recall, and F1-score of 98.73%, which is superior to the EfficientNet-B7 base model. The findings of this study demonstrate the potential of hybrid models to improve accuracy in lung cancer classification. The utilization of hybrid models has the potential to contribute significantly to the beginning diagnosis and appropriate Lung Cancer Therapies. Future research will focus on improving the model through the application of image segmentation techniques and expanding the scope of classification to other types of lung cancer. Optimization of the hybrid model architecture using novel techniques such as the attention mechanism or transfer learning will be conducted to improve the efficiency and accuracy of the model. Additionally, a system that can be integrated into clinical practice will be develope
Transliteration of Latin Letters to Bali Characters Based on Unicode for Mobile Devices using Finite State Automata and Levenshtein Distance
No full textThe preservation of Balinese script writing has been pursued by the local government with the issuance of the Bali Province Regional Regulation Number 1 of 2018 concerning Balinese Language, Script, and Literature. However, the use of Balinese script in daily life is declining, especially among generation Z, most of whom find Balinese language difficult to use and have never been taught it. Technological advances, particularly smartphone technology, can play an important role in shaping generation Z's habits, values, and social interaction patterns. This research uses the Finite State Automata (FSA) method to convert Latin letters to the Balinese script Unicode standard, following the Balinese script writing rules. FSA governs transliteration behavior by using the working principles of State, Event, and Action. Besides transliterating sentences typed by users, the application produced by this research also displays Balinese script words related to the words typed by users using the Levenshtein Distance method. The ‘related words’ feature allows users to know more about Balinese script than just the typed word. From the test results conducted through two different test cases, the first test case tested the application's ability to transliterate words/sentences typed by users without selecting words from the application's suggestions. The results showed that of the 50 words tested, 39 were correctly transliterated. The second test case tested the app's ability when the user selects a word from the suggestions given by the app. The result shows that out of 50 words tested, 43 transliteration data are correct, with the total accuracy of both test cases being 82%
Machine Learning to Predict Student Satisfaction Level Using KNN Method and Naive Bayes Method
No full textThis research aims to apply machine learning techniques in classifying student satisfaction levels at the Faculty of Science and Technology, using the K-Nearest Neighbors (KNN) and Naive Bayes methods. This method was chosen because of its ability to manage classification data and provide accurate predictions regarding student satisfaction with the faculty. It is hoped that this research will provide a deeper understanding of the factors that influence student satisfaction as well as the potential for developing a better evaluation system in the future. This research was carried out through structured stages, starting from selecting the right data to collect relevant information, designing the model by applying the KNN and Naive Bayes methods, to evaluating the performance of the model being built. The data used consisted of 110 student data, where the classification results showed that 104 students were satisfied and 6 other students were dissatisfied with the faculty. The evaluation process produced excellent accuracy, with the Test and Score results and confusion matrix showing an accuracy level exceeding 90%. In conclusion, this research succeeded in showing that the KNN and Naive Bayes methods were effective in classifying the level of student satisfaction at the Faculty of Science and Technology. The results obtained confirm that both methods are reliable in managing and analyzing student satisfaction data efficiently, and provide valuable insights for educational institutions to improve student services and experiences in the futur
Real-Time Monitoring of Photovoltaic Panel Using Node-RED
No full textThis research aims to design and implement an Internet of Things (IoT)-based monitoring system for Photovoltaic (PV) panels using Node-RED. The system can monitor critical parameters such as voltage, current, power, and the electrical energy produced by the PV panels in real-time. The data obtained from the PV panels is sent to a Node-RED server and visualized in the form of indicators and graphs on a dashboard. Statistical analysis calculates the daily average power and total energy produced. The results show that the proposed system can enhance monitoring efficiency and significantly benefit PV system maintenance and management. Users can quickly identify and address issues that may arise, such as panel performance degradation or system disruptions. Energy analysis and maintenance planning can be carried out by collecting historical data. This research supports the broader renewable energy development and provides an effective real-time PV system monitoring solution
Master Stockist Customer Segmentation Using RFM Model and Self-Organizing Maps Algorithm
No full textMaster Stockist PT SNS 21 Bali struggles to identify member performance based on purchasing behavior because the applicable system only records transactions and stock of goods without providing insight into customers. Customer segmentation can be carried out to identify and understand differences in customer purchasing behavior. Therefore, this study aims to determine customer segmentation using the RFM (Recency, Frequency, Monetary) model and the Self-Organizing Maps (SOM) algorithm. Segmentation development uses the CRISP-DM (Cross-Industry Standard Process for Data Mining) approach. The RFM model numerically represents customer behavior through three variables, while the Self-Organizing Maps algorithm groups customers into segments with similar characteristics. In this research, the best SOM parameters are 750 iterations, learning rate 0.5, radius 0.5, and grid size 1x3, resulting in 3 clusters with a Silhouette Score of 0.647608 and a Davies-Bouldin Index of 0.536503. Cluster 1 consists of 226 new customers with low RFM values who need encouragement to be more active. Cluster 2, comprising seven members, has low recency, high frequency, and high monetary values, representing loyal customers who need to be retained. Cluster 3 consists of 239 inactive customers with high recency, low frequency, and low monetary values, requiring a reactivation strategy
Clustering Analysis of Cadet Profiles Using Age, Recency, Frequency and Monetary Methods Using K-Means and K-Medoids Algorithms
No full textBanten Maritime Polytechnic is a new academic school established in 2019 so that the formulation of data management is still being sought to be suitable and optimal, there are many obstacles if the data is not managed properly, starting from the recruitment of prospective cadets in taking sailor competency training such as not optimal socialization. According to data from the 2021 Transportation Human Resource Development Agency, it explains that there are still few enthusiasts, especially at the Banten Maritime Polytechnic. The purpose of this study is to analyze the profile of cadets in taking sailor competency training using the age, recency, frequency and monetary methods in categorizing data and clustering with the k-means and k-medoids algorithms so that the data can be used for cadet services and related parties in the Banten Maritime Polytechnic database. This analysis can also be used for mapping in recruiting prospective cadets in taking sailor competency training so that they can see opportunities and optimize target markets. This research was conducted in 2023 based on the latest data on the 2022-2023 academic year cadet profile at the Banten Maritime Polytechnic. The results of this analysis data can be used for cadets who have not graduated and have graduated in finding work partners and channeling cadets to the shipping industry. So it is very important to manage and cluster cadet profile data in taking this sailor competency training. The use of the K-means and K-medoids algorithms helps in compiling data groupings that have large data. It works by looking at the number of small groups or groups whose numbers are represented by the variable K. To be able to group the existing data, the K-means algorithm runs iteratively from each existing data point to the K group that has been created. The results of the study are cadet profile grouping data that can be managed again for strategies and management formulations at the Banten Maritime Polytechnic, especially in increasing the recruitment of prospective cadets in taking sailor competency training
Forward Selection as a Feature Selection Method in the SVM Kernel for Student Graduation Data
No full textIn the era of information technology development, accurate graduation predictions are important to improve the quality of higher education in Indonesia. This research evaluates the effectiveness of Support Vector Machine (SVM) with various kernels, including Radial Basis Function (RBF), linear, and polynomial, as well as the application of FS as an optimization method. The dataset used consists of student graduation data which includes nine independent attributes and one label. This research aims to increase the accuracy of student graduation predictions using the SVM method which is optimized through Forward Selection (FS). The SVM method is applied using 10-fold cross validation to predict on-time graduation. The results show that the combination of SVM and FS improves prediction accuracy significantly. The SVM model with an RBF kernel optimized with FS achieved the highest accuracy of 87.06% and recall of 53.68%, showing increased sensitivity in identifying student graduation cases compared to SVM without FS. Although there is a trade-off between precision and recall, the model optimized with FS shows better performance overall. This research contributes to the development of a more efficient graduation prediction method, which can help universities in planning strategies to improve academic quality. Further studies are recommended to overcome weaknesses in the recall value by using other optimization methods or combinations of other optimization algorithm
Analysis of Factors That Affects COVID-19 Vaccination on Countries Worldwide
No full textDespite the urgency of vaccination against COVID-19 worldwide, each country has different levels of vaccination rate which lead to different success rates. While several past studies have shown what factors affect a country’s vaccination rating from past epidemics, there are no correlation studies done on factors to COVID vaccination rates, with several media and institutes forming theories, with New York Times stating it’s GDP per Capita, and National Health Institute postulating literacy and other various factors, while none those two showing correlation studies of the factors as the proof. With values ranging from -1 to 1, results showed among six factors ranging from 0.51 to 0.64 for four factors showing that of six factors listed in this study, meaning they are moderately strongly related with the vaccination rate, with one having a value of 0.14, meaning it’s weakly related, another with value of -0.58, indicating strongly unrelated with vaccination rate