Sinkron : jurnal dan penelitian teknik informatika
Not a member yet
1261 research outputs found
Sort by
Enhancing Supervised Learning through Empirical Enrichment Using Style Transfer Generative Datasets
No full textAn innovative strategy for improving supervised learning by utilizing empirically enriched datasets through the application of generative style transfer techniques. Within the realm of artificial intelligence, supervised learning has emerged as a significant domain. However, the challenge of acquiring datasets that are both representative and diverse persists. To tackle this issue, this research integrates the notion of style transfer to broaden the range of data accessible for supervised learning models. This method employs the style transfer process to generate diverse style variations within the existing data. Incorporating various image variations enhances the dataset and enables the model to gain a deeper comprehension of the image's content. Experiments were performed utilizing a conventional dataset that was enhanced using a style transfer technique and subsequently inputted into a supervised learning model. The results demonstrate substantial enhancements in model performance, particularly in terms of its ability to generalize to new test data. This confirms the efficacy of this approach in enhancing the quality of supervised learning. These findings emphasize the significant potential of employing style transfer in dataset enrichment to improve and intensify model comprehension in managed learning scenarios, as well as its implications in the advancement of artificial intelligence technologies that are more flexible and capable of adjusting to various visual scenarios
Implementation of App Engine and Cloud Storage as REST API on Smart Farm Application
No full textSmart Farm is an agricultural application that uses machine learning and cloud computing technology to improve efficiency in the farming process. Technological advancement and sustainable agriculture are two essential aspects of supporting global food security. This research investigates the implementation of App Engine and Cloud Storage in developing REST API in Smart Farm applications. By utilizing cloud computing technology, such as App Engine, and cloud storage, such as Cloud Storage, we can create efficient solutions to monitor and manage agriculture better. This research implements an App Engine and Cloud Storage to develop a REST API that allows Smart Farm application users to access data and control farming devices efficiently. The authors designed, developed, and tested this system to ensure optimal performance and reliability in agricultural data collection and distribution. This method has several significant advantages. First, App Engine allows for easy scalability, ensuring the system can handle increased data demand without disruption. Secondly, Cloud Storage provides secure and scalable storage for agricultural data, which can be accessed from anywhere. This provides easy and quick access to critical data for farmers. Moreover, the use of cloud technology also reduces infrastructure and maintenance costs. The developed system integrates the App Engine and Cloud Storage with the Smart Farm application. The App Engine is a processing engine that receives user requests via the REST API, processes the required data, and provides appropriate responses. Like image data, farm data is stored and managed on Cloud Storage. Users can access this data through the Smart Farm app or other devices, enabling better farming monitoring and decision-making
Prediction of Stunting in Toddlers Using Bagging and Random Forest Algorithms
No full textStunting is a condition of failure to thrive in toddlers. This is caused by lack of nutrition over a long period of time, exposure to repeated infections, and lack of stimulation. This malnutrition condition is influenced by the mother's health during pregnancy, the health status of adolescents, as well as the economy and culture and the environment, such as sanitation and access to health services. To find out predictions of stunting, currently we still use a common method, namely Secondary Data Analysis, namely by conducting surveys and research to collect data regarding stunting. This data includes risk factors related to stunting, such as maternal nutritional status, child nutritional intake, access to health services, sanitation, and other socioeconomic factors. This secondary data analysis can provide an overview of the prevalence of stunting and the contributing factors. To overcome this, the right solution is needed, one solution that can be used is data mining techniques, where data mining can be used to carry out analysis and predictions for the future, and provide useful information for business or health needs. Based on this analysis, this research will use the Bagging method and Random Forest Algorithm to obtain the accuracy level of stunting predictions in toddlers. Bagging or Bootstrap Aggregation is an ensemble method that can improve classification by randomly combining classifications on the training dataset which can reduce variation and avoid overfitting. Random Forest is a powerful algorithm in machine learning that combines decisions from many independent decision trees to improve prediction performance and model stability. By combining the Bagging method and the Random Forest algorithm, it is hoped that it will be able to provide better stunting prediction results in toddlers. This research uses a dataset with a total of 10,001 data records, 7 attributes and 1 attribute class. Based on the test results using the Bagging method and the Random Forest algorithm in this research, the results obtained were class precision yes 91.72%, class recall yes 98.84%, class precision no 93.55%, class recall no 65.28%, and accuracy of 91.98%
Diabetes Disease Detection Classification Using Light Gradient Boosting (LightGBM) With Hyperparameter Tuning
No full textDiabetes is a condition caused by an imbalance between the need for insulin in the body and insufficient insulin production by the pancreas, causing an increase in blood sugar concentration. This study aims to find the best classification performance on diabetes datasets with the LightGBM method. The dataset used consists of 768 rows and 9 columns, with target values of 0 and 1. In this study, resampling is applied to overcome data imbalance using SMOTE and perform hyperparameter optimization. Model evaluation is performed using confusion matrix and various metrics such as accuracy, recall, precision and f1-score. This research conducted several tests. In hyperparameter optimization tests using GridSearchCV and RandomSearchCV, the LightGBM method showed good performance. In tests that apply data resampling, the LightGBM method achieves the highest accuracy, namely the LightGBM method with GridSearchCV optimization with the highest accuracy reaching 84%, while LightGBM with RandomSearchCV optimization reaches 82% accuracy
Integration of Artificial Intelligence in Facial Recognition Systems for Software Security
No full textFacial recognition technology, a cornerstone in modern software security, has seen significant advancements through the integration of Artificial Intelligence (AI). This research focuses on enhancing facial recognition systems by incorporating sophisticated machine learning algorithms and deep neural networks. By doing so, the goal is to increase the accuracy and reliability of these systems in security applications. The study uses a variety of facial datasets to train AI models that are adept at extracting facial features and recognizing patterns. These models are subjected to rigorous testing to evaluate their performance in terms of identification accuracy, processing speed, and adaptability to different environmental conditions. One of the key challenges addressed in the research is the system's vulnerability to errors and potential misuse. Ethical considerations and privacy concerns are at the forefront of the study. The research highlights the importance of designing AI-based facial recognition systems that respect user privacy and are resistant to biases, thus fostering trust and acceptance among users. The results of the study show a marked improvement in system performance, demonstrating enhanced recognition accuracy and speed, while maintaining robustness across different conditions. By offering practical recommendations for the development of secure, ethical, and privacy-aware facial recognition systems, this research contributes valuable insights into the integration of AI in software security. It underscores the importance of continuous innovation and ethical responsibility in the deployment of facial recognition technologies, shaping the future landscape of technological security measure
Sentiment Analysis of Public Responses on Social Media to Satire Joke Using Naive Bayes and KNN
No full textThis study examines the use of Satire Joke as a humorous communication style in conveying criticism of the government through social media. Satire Joke is often used to depict the government's inability to address important social issues, such as slow bureaucratic processes and unfulfilled political promises. The aim of this research is to analyze public sentiment towards Satire Joke expressed on the YouTube social media platform. The methods used in this study are Naïve Bayes and K-Nearest Neighbors (KNN) due to their effectiveness in data classification. The results of this study are expected to help gain an understanding of social issues for the community and public knowledge. This research is also expected to contribute to the development of sentiment analysis methods in the future. The analysis results show that 400 data have neutral sentiment, 850 data have negative sentiment, and 947 data have positive sentiment. Based on testing, both Naive Bayes and KNN methods show good performance. The Naive Bayes method achieved the best accuracy of 90.29%, while the KNN method achieved an accuracy of 60.75%
Decision Support System for Selecting Online Teaching Methods Using the Fuzzy MCDM Algorithm
No full textThe global pandemic that has hit the world recently has forced educational institutions to adopt online teaching methods. However, choosing an effective online teaching method is a major challenge. This research develops a Decision Support System (DSS) that uses the Fuzzy Multi-Criteria Decision Making (FMCDM) Algorithm to select the best online teaching method. This system is designed to assist decision making in educational institutions by considering various criteria such as learning effectiveness, technology affordability, ease of use, and user satisfaction. This research uses data collection methods that involve surveys from lecturers and students to obtain their preferences and experiences with various online teaching platforms. The data collected is then processed using the FMCDM model to evaluate and rank teaching methods based on predetermined criteria. Fuzzy systems are used to overcome uncertainty and subjectivity in criteria assessment. The results of this research show that the system developed is able to effectively assess and rank various online teaching methods. From the analysis carried out, interactive teaching methods using videos and real-time quizzes received the highest ranking based on predetermined criteria. This suggests that the combination of engaging visual content and high interactivity is highly valued in online teaching context
Optimization Model for Relief Distribution After Flood Disaster
No full textLogistics planning is critical and a key component in meeting initial emergency needs in the aftermath of a disaster. The rapid and efficient distribution of logistical aid becomes critically important. In such situations, the construction of temporary depots in strategic locations and the determination of optimal distribution routes play an important role in ensuring that logistics aid can be distributed to the affected areas evenly. In this study, the Multi Depot Vehicle Routing Problem (MDVRP) is used which aims to minimize the total cost of distributing logistics aid which includes shipping costs, vehicle usage costs, temporary depot construction costs, and vehicle travel costs from distribution centers to temporary depots, while still meeting constraints such as logistics aid demand, vehicle capacity, area visits, maximum mileage, and depot construction. This model uses two types of vehicles where vehicle is tasked with carrying logistics aid from the distribution center to the temporary depot and vehicle is tasked with delivering logistics aid directly to the point of demand
Integration of AHP and Modified VIKOR Method to Select the Optimum Destination Route
No full textOne common approach to rating options is group decision making using many criteria. Here, we use the same criteria to evaluate each option. Sometimes, decision makers are faced with some situations where they have to choose from a set of alternatives that have several different criteria. Thus, the decision maker cannot use a common method. Therefore, in this research, a modification to a method is carried out. To address the issue of developing alternate routes to Medan City's historical tourism attractions, the AHP and VIKOR approaches have been suggested. When considering options with both specific and broad requirements, this study adapts the VIKOR technique to find a workable solution. In order to demonstrate the suggested model's use and evaluate the efficacy of this approach change, this study offers numerical examples based on case studies. The findings demonstrate that the revised approach is both practical and efficient
ChatBot-based Bus Ticket Booking Prototype Using WhatsApp
No full textCustomer relationship management (CRM) is the most critical part of any business's operations. To deploy CRM at PO Harapan Jaya, a ticket ordering system is required to make things easy for customers. Researchers constructed a chatbot prototype using the WhatsApp app with the goal of making it easier and more efficient for PO Harapan Jaya customers to buy bus tickets. The eXtreme Programming (XP) approach is a strategy for creating a ticket-booking chatbot prototype using WhatsApp. In developing the WhatsApp chatbot prototype for this research, a CRM system is proposed that can collect and manage information entered by customers, ticket purchase history, and produce proof of ticket reservations via WhatsApp chat, which will then be shown to the PO Harapan Jaya admin to obtain bus tickets ordered by customers. Researchers conducted a black box test on the prototype. This study intends to demonstrate how deploying chatbots in the WhatsApp application may speed up the bus ticket purchase process, increase customer service quality, and assist PO Harapan Jaya in optimizing CRM tactics. Based on the extensive programming processes, including planning, design, coding, and testing, it is possible to determine that the final WhatsApp chatbot will work properly and may be used by users to buy PO Harapan Jaya bus tickets. Customers may book tickets, examine departure schedules, and contact the PO Harapan Jaya admin if they have any issues when ordering tickets