Sinkron : jurnal dan penelitian teknik informatika
Not a member yet
1261 research outputs found
Sort by
Comparison of Performance of K-Nearest Neighbors and Neural Network Algorithm in Bitcoin Price Prediction
No full textThis research evaluates and compares the performance of two prediction methods, namely K-Nearest Neighbors (K-NN) and Neural Network, in the context of Bitcoin price prediction. Historical Bitcoin price data is used as input to train and test both algorithms. Experimental results show that the K-NN algorithm produces a Root Mean Square Error (RSME) of 389,770 and a Mean Absolute Error (MAE) of 89,261, while the Neural Network has an RSME of 614,825 and an MAE of 284,190. Performance comparison analysis shows that, on this dataset, K-NN has better performance in predicting Bitcoin prices compared to Neural Network. These findings provide important insights for the selection of crypto asset price prediction models, especially Bitcoin, in financial and investment environment
Comparison of Naïve Bayes and SVM in Sentiment Analysis of Product Reviews on Marketplaces
No full textAt this time more and more people are switching to shopping online in existing marketplaces such as Shopee. Marketplaces provide various advantages and disadvantages to customers such as lower costs and goods sent not according to orders. Product reviews from customers greatly affect the sales level of business people so that sentiment analysis is carried out. The importance of conducting sentiment analysis of product reviews in the marketplace is to add an overview of how the product is received by users. This research uses Naïve Bayes and SVM algorithms for sentiment analysis of beauty care product review datasets obtained from Shopee scraping results. This research implements k fold cross validation for data splitting process of 10 folds. The Naïve Bayes algorithm obtained the highest accuracy value of 85.53% on fold 2 and the lowest accuracy value of 77.16% on fold 3. While the SVM algorithm obtained the highest accuracy value of 88.58% on fold 2 and the lowest accuracy value of 82.99% on fold 7. With this it is stated that SVM can work better for sentiment analysis of beauty care product reviews on the Shopee marketplace because it gets a higher average accuracy value of 86.14% compared to the Naïve Bayes algorithm
An In-Depth Analysis of SIMPKB: Revealing Performance Tests and Efficiency from a User Experience
No full textThis study comprehensively analyzes the performance and usability of the SIMPKB website in the context of teacher professional development. This research carries a qualitative descriptive approach with the aim of deeply understanding the performance and usability of the SIMPKB website. This research consists of two complementary stages, the first involves performance testing using GT Metrix software, and the second phase focuses on in-depth interviews with 5 driving teachers in Kabupaten Jember by applying the concept of the Five Dimensions of Usability (5E) model. Through performance testing using GT Metrix and 5E interviews with driving teachers, significant findings have been revealed. Although SIMPKB shows relatively good response speeds, there are areas of improvement that can be improved, especially in terms of loading times and Largest Contentful Paint (LCP). The 5E evaluation of the mobilizing teacher provides an in-depth perspective on the effectiveness, efficiency, engagement, errors, and ease of learning on the platform. The test and interview results complement each other, providing a holistic picture of SIMPKB's condition and potential improvement. Improvement recommendations, which involve improving response speed and improving usability, can be a foothold for improving the user experience. Further research is recommended to explore optimizing technical performance, implementing more intuitive interface designs, and evaluating the impact of implementing improvements on user effectiveness. By adopting these recommendations, SIMPKB can continue to develop as an effective, efficient, and user-friendly platform in supporting teacher professional development
Feasibility Analysis of Bengkel Koding Website Using Black Box Testing and Boundary Value Analysis
No full textIn an era of rapid technological development, application development has become common, especially in coding. However, most websites do not give appropriate assignments and instructors to help improve coding skills. Because of this, the Bengkel Koding of Dian Nuswantoro University Semarang is a solution to improving the quality of coding learning. This research aims to identify the shortcomings in the website and ensure that the website functions as expected by the users. By testing the application like this, researchers can know which problems can affect the user experience. This research uses one of the frequently used tests, namely Black Box testing. The objective is to verify that the system's functions, inputs, and outputs align with the specified requirements. In addition to the Black Box method, this research uses a technique called Boundary Value Analysis. This technique is to identify errors or bugs that can affect the user experience by focusing on the input value boundary. The test results will use a quality ratio that will determine whether or not the system is suitable for use by users. Through 30 test cases, most website functions have been tested properly, with the feasibility level reaching 83.333%. Nonetheless, five errors or bugs were still found, emphasizing the need for further improvement. The results of this study provide valuable insights into improving the quality and convenience of users in accessing the Bengkel Koding website
Enhancing Cable News Network Comprehension: Text Rank Integrated Natural Language Processing Summary Algorithm
No full textIn the online news space, timely content delivery has become essential due to the unavoidable information overload. This study investigates the use of Python-based text summarizing techniques on news sites, promoting the combination of Natural Language Processing approaches with the Text Rank summarization algorithm. The primary objective is to deliver automatic news article summaries while preserving pertinent information, this is confirmed by means of experimental testing. This study uses the Text Rank technique on a news platform to enhance summaries' readability and information absorption capacity. To test the Text Rank algorithm's capacity to provide enlightening summaries, two news stories from the Cable News Network were chosen for the experiment. The word "Trump" obtained the highest score of 16.52 when sentence scores were calculated using the Text Rank algorithm. "Former" came in second with a score of 1.95, "McCarthy" was third with a score of 1.31, and "President" and "Republican" were each awarded a score of 1.03. Furthermore, the terms "CNN" and "Establishment" received scores of 0.79 and 0.58, respectively, for "DeSantis" and "Endorsements." Reader accessibility and convenience can be improved by using a news summary algorithm on a Python-based platform to swiftly retrieve important information. This research emphasizes the critical role that summary algorithm technology plays in enabling efficient and easily accessible information consumption in the digital age, in addition to creating automated tools for news summaries
Prediction of Stunting in Toddlers Combining the Naive Bayes Method and the C4.5 Algorithm
No full textResearch conducted to predict the incidence of stunting in toddlers, using data mining methods such as Naive Bayes and the C4.5 algorithm has been applied to analyze health data. The main aim of this research is to develop a predictive model that can identify toddlers who are at high risk of stunting, based on variables that have been collected from medical records and health surveys. The use of the Naive Bayes and C4.5 methods in this research aims to compare the effectiveness of the two methods in dealing with complex and unbalanced classification problems. This research involves a series of crucial stages starting from data selection, data pre-processing, data mining model design, data mining model testing, to method evaluation. In this study, the sample used consisted of 200 toddlers, of which 159 were diagnosed as having stunting and 41 others were not. The classification results show significant effectiveness in both methods used. The accuracy results of both methods are very encouraging, with both methods showing success rates of more than 90%. This shows that both Naive Bayes and C4.5 are very effective in identifying patterns related to the risk of stunting among toddlers. These highly accurate results not only demonstrate the power of data mining techniques in the field of public health but also provide insights that health practitioners can use to intervene earlier in at-risk populations
Ontology-based Food Menu Recommender System for Pregnant Women Using SWRL Rules
No full textPregnancy is a crucial period in a woman's life because her body must prepare and support the growth and development of the fetus. During pregnancy nutritional needs will increase. Lack of nutritional intake during pregnancy can cause serious health problems, one of which is anemia. However, excess nutrition during pregnancy also has a negative impact on pregnant women. Therefore, a recommender system is required to provide food menu recommendations according to the daily nutritional needs of pregnant women. Currently, there has been a lot of research on ontology-based food recommender systems that can provide food recommendations to users, but there is no research that specifically provides food menu recommendations that suit the needs of pregnant women. Therefore, in this research, we propose an ontology-based food menu recommender system using SWRL (Semantic Web Rule Language) rules for pregnant women. In this food menu recommender system, ontology is used to represent food knowledge and its nutritional content, and SWRL rules are used to reason logical rules in the ontology to determine the appropriate food menu for pregnant women. This recommender system also considers diseases and allergies that pregnant women have so that it can provide food menu recommendations that are more suitable for users. From 15 data samples from pregnant women, the system provides 75 food menu recommendations for pregnant women. Based on the validation results that have been carried out, the precision value is 0.986, the recall is 1, and the F1-score is 0.992
Implementation Transfer Learning on Convolutional Neural Network for Tubercolosis Classification
No full textTuberculosis (TB) is an infectious disease that can have serious effects on the lungs and is among the top 10 causes of death worldwide. This disease is caused by the transmission of Mycobacterium tuberculosis bacteria through the air when coughing or sneezing. Without treatment, pulmonary tuberculosis can result in permanent lung damage and can be life-threatening. Accurate and early diagnosis is crucial for effective treatment and control of the disease.The challenge lies in the accurate classification of tuberculosis from lung images, which is essential for timely diagnosis and treatment. Traditional diagnostic methods can be time-consuming and sometimes lack precision. To address this issue, this research aims to achieve high accuracy in classifying tuberculosis using the Convolutional Neural Network (CNN) algorithm through transfer learning methods. By utilizing visual images of tuberculosis-affected and normal lungs, we propose a solution that leverages advanced deep learning techniques to enhance diagnostic accuracy. This approach not only expedites the diagnostic process but also improves the reliability of tuberculosis detection, ultimately contributing to better patient outcomes and more effective disease management. The dataset applied consists of two labels: tuberculosis and normal. This dataset contains 4200 lung images of individuals with tuberculosis and normal lungs. By applying the transfer learning method, Transfer learning is a machine learning method where a pre-trained model is used as the starting point for a new, related task. it was found that the ResNet50 model achieved the highest accuracy at 99%, followed by InceptionV3 at 97%, and lastly, DenseNet121 at 91%
Robust Regression on Simple Housing Environmental Performance Measurement
No full textRobust regression in residential environmental performance measurement can provide significant benefits. This method can help identify and overcome uncertainties in measurement results so that decision making related to environmental protection and management can be done more accurately. This type of research is quantitative with a descriptive approach. The results of this study indicate that the results of the M-estimate robust regression estimation on simple housing environmental performance obtained the M-estimate robust regression equation, namely ŷ = 15.562+0.476X1-0.453X2+0.222X3-0.427X4. Based on the p-value test results, it shows that population attributes, house attributes, energy consumption volume and utilities and services have a significant effect on environmental performance. The results of the hypothesis testing of the M-estimation robust regression method on the environmental performance of simple housing obtained that the variables of population attributes and energy consumption volume have a positive and significant effect on environmental performance while the variables of house attributes and utilities and services have a negative and significant effect on environmental performance
Deployment of Web-Based YOLO for CT Scan Kidney Stone Detection
No full textThis research aims to develop a kidney stone object detection system using machine learning techniques like YOLO and object detection, integrated into a Flask-based web interface to support early diagnosis by medical professionals. The trained model demonstrates strong pattern learning capabilities. Evaluation of the public dataset model reveals an average mean Average Precision (mAP) of 0.9698 for 'kidney stone' labels. This detection model exhibits high performance with an accuracy rate of 96.33%, precision of 96.98%, recall of 99.23%, and an F1-score of 98.1%. Clinical data evaluation shows that the YOLOv5-based detection system performs exceptionally well, with an average mAP of 0.9571, accuracy of 93.06%, precision of 95.71%, recall of 97.1%, and F1-score of 96.49%, indicating the model's capability to detect kidney stones with high precision and accuracy. Thus, both the evaluation on the public dataset and clinical dataset performance support accurate diagnosis processes and further treatment planning. Moreover, this research advances to the stage where the detection model can be directly utilized through implementation via Flask web deployment