JUTI: Jurnal Ilmiah Teknologi Informasi
Not a member yet
407 research outputs found
Sort by
MACHINE LEARNING JOURNAL ARTICLE RECOMMENDATION SYSTEM USING CONTENT BASED FILTERING
Full text linkIndonesia is a country that hasn’t studied much about artificial intelligence. This has resulted in a small number of publications related to that field including areas within such as machine learning. For that reason, it caused difficulties in finding relevant journal articles. The purpose of this study is to know the performance of the Content Based Filtering method in providing machine learning journal article recommendations. The research procedure used is CRISP-DM with algorithms used are TF-IDF and Cosine Similarity. The dataset used consists of 100 machine learning journal articles. Based on the research that has been done, it’s concluded that the performance of the Content Based Filtering method in providing machine learning journal article recommendations as measured using the precision evaluation matrix showed a score of 76%, which means the result is quite good. However, the model couldn’t be used properly for some data due to the small number of datasets which affects the limited recommendations.
MOBILE-BASED ONLINE QUEUE APPLICATION DEVELOPMENT AT GRIBIG PUBLIC HEALTH CENTER IN REALTIME
Full text linkPuskesmas is one of the health service facilities that organizes community and individual health efforts at level one. Gribig Health Center is one of the health centers in the city of Malang, East Java, which has several health services. To get services from the Puskesmas, each patient is required to register and complete the required files for further processing by the administration. However, the imbalance between the number of patients and the availability of services is the cause of queues. This study aims to create a mobile-based queuing application at the Gribig Health Center in real-time. The ar-chitectural concept used in developing applications is client-server. The queuing method used in the system is a combina-tion of FCFS (First Come First Served) and PS (priority service) methods. In system development, the development method used in this research is the waterfall method. For system testing, the author uses the Black Box Testing method to ensure that all application functionality is appropriate. The purpose of developing this application is to make it easier for pa-tients to get queue numbers for Gribig Health Center services anywhere and anytime, make it easier for patients to make registration bookings for other days in advance, exchange queue numbers, notifications when their turn is approaching, find out the estimated time to get service, and queue information. up-to-date for each service at the Puskesmas. The results of this study are successful in developing an online queuing application at the Gribig Health Center in real-time by utiliz-ing the QR Code to verify the queue and there is also a notification feature as a patient reminder
COAL DEMAND PREDICTION MODEL USING MACHINE LEARNING METHODS
Full text linkForecasting coal demand needs is important to minimize operational costs. Forecasting will help companies determine the right amount and time to order coal from suppliers. Research on coal forecasting in Indonesia generally uses a statistical approach and has not analyzed the performance of other forecasting models. This research aims to forecast coal demand using statistical and machine learning methods, namely ARIMA, Exponential Smoothing, Support Vector Regression (SVR), Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM). The evaluation methods used to analyze forecasting performance are Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). The new coal demand data used is 1097 daily data taken from January 2021 to December 2022 in the form of a timeseries and is stationary which has been tested using Augmented Dickey-Fuller (ADF). The test results show that the ARIMA model has MAPE value of 5.11%, MAE 2.91 and R-Square 0.925, Exponential Smoothing MAPE 1.07%, MAE 0.55 and R-Square 0.997, SVR with MAPE value of 5.48%, MAE 3.16 and R-Square 0.88, RNN with MAPE value of 5.19%, MAE 2.91 and R-Square 0.896, LSTM with MAPE value of 4.83%, MAE 2.84 and R-Square 0.897. From the test results it was found that exponential smoothing had the smallest error values among the other models. With forecasting results that have a small error rate, it can help management in making decisions to minimize costs in coal ordering and warehouse management
DEVELOPMENT OF A MODEL TO EVALUATE USERS\u27 TECHNOLOGY READINESS AND ACCEPTANCE IN USING THE SELF-CHECK-IN KIOSK SERVICE AT SOEKARNO-HATTA INTERNATIONAL AIRPORT
Full text linkThe self-check-in kiosk is one of the digital technologies used by the aviation industry to help passengers check in on passenger flights independently and efficiently without the need for a conventional check-in counter at the airport. However, the phenomenon on the ground indicates that many users have not yet used the service. As a result, the check-in area in some of the flight masks often has a long wait. Studies conducted by several airports in campsites such as Malaysia, South Africa, and Switzerland show that self-check-in kiosks do not meet the echoes of users. The same thing happened at Indonesian airports, where the use of self-check-in kiosks was still below 20% of total passenger traffic in 2022–2023. The study introduces the User Experience Technology Readiness and Acceptance Model (UX TRAM), which is used to evaluate user readiness and acceptance of the application of new technologies in the airport environment. The Partial Least Squares Structural Equation Modeling (PLS-SEM) method is used to analyze the research model and the proposed hypothesis. Based on the results of the test of significance and relevance of the relationship in this study, the structural model proposed by the majority is of significant value, except for the variables Innovativeness and Insecurity versus Perceived Ease of Use. Based on the results of the test of the hypothesis carried out, out of 15 hypotheses tested, there are 13 accepted and 2 rejected hypotheses related to the readiness and acceptance of users in the use of new technology on the Self-Check-in Kiosk service at Soekarno-Hatta International Airport. The results of this study show that the proposed research model has varying explanatory strengths (near moderate to substantial/high) as well as predictive strengths that offer better predictable performance.
ANALYSIS OF USER EXPERIENCE OF DANA E-WALLET USING USER EXPERIENCE QUOSTIONNAIRE (UEQ) AND UX HONEYCOMB
Full text linkIn the era Society 5.0, economic growth and digital products have mushroomed, leading to digital transaction product innovation becoming a daily necessity for society. Easy, cashless payment are made through an E-wallet application with various options, including the DANA application. However, in several reviews, user have complained about the performance and issues with the DANA application such as, the display is unclear when logging into the application, and it loads slowly when entering the amount of money. This research was conducted to evaluate DANA application users and provide recommendation for improvements and enhancements to enhance the DANA application user experience. From the result of the value proposition on the implementation of the UX Honeycomb method in the DANA application between the people of Palembang City, users generally agree that DANA deserves a good above-average rating. Meanwhile, the User Experience Questionnaire (UEQ) method show that the DANA application received positive scores on the attractive, perspicuity, dependability, efficiency, and stimulation variables, but it obtained a negative score on the novelty variable. Thus, the DANA application need to design and create more creative, intentional, and innovative products
CLASSIFICATION OF LUNG AND COLON CANCER TISSUES USING HYBRID CONVOLUTIONAL NEURAL NETWORKS
Full text linkColon and lung cancers are two highly lethal kinds of cancer which can often coexist and pose a new challenge for accurate diagnosis. While research often concentrates on detecting a single cancer in a specific organ, this study proposes an innovative machine-learning approach to identify both colon and lung cancers. The objective is to create a hybrid machine learning classification model to enhance diagnostic precision. The LC25000 dataset comprises 25,000 color histopathological image samples of lung and colon cell tissues, indicating the presence or absence of cancer (adenocarcinoma). Image features are extracted using the pre-trained VGG-16 model. The cancer type is identified through three machine learning classification algorithms: Stochastic Gradient Descent (SGD), Random Forest (RF), and K-Nearest Neighbor (KNN). The model\u27s evaluation employed a 10-fold cross-validation technique, with CNN-SGD exhibiting the highest performance based on evaluation metrics. On a scale of 0 to 100, it scored 99.8 for Area Under Curve (AUC) and 98.88 for Classification Accuracy (CA). CNN-RF, a model with performance closely following CNN-SGD, demonstrates training times 58.3 seconds faster than CNN-SGD. Meanwhile, CNN-KNN ranks last among the models evaluated in this study based on its F1, recall, AUC, and CA scores
UAV LAND COVER CLASSIFICATION USING CONVOLUTIONAL NEURAL NETWORK FEATURE MAP WITH A COMBINATION OF MACHINE LEARNING
Full text linkIn geographic analysis, land cover plays an important role in everything from environmental analysis to sustainable planning methods and physical geography studies. The Indonesian National Standard (SNI 7645:2014) classifies vegetation analysis based on density. There are four categories of vegetation density index: non-vegetation, bare, medium, and high. Technically, vegetation data can be obtained through remote sensing. Satellite and UAV data are two types of data used in remote sensing to collect information. This research will analyze land cover based on vegetation density information that can be collected through remote sensing. Based on vegetation density information from remote sensing, the information can help in land processing, Land Cover Classification is carried out based on vegetation density. Convolutional neural networks (CNN) have been trained extensively to apply their properties to land cover classification. This research will evaluate features extracted from Convolutional Neural Networks (ResNet 50, Inception-V3, DenseNet 121) which have previously been trained and continued with Decision Tree algorithms, Random Forest, Support Vector Machine and eXtreme Gradient Boosting to perform classification. From the comparison results of classification tests between machine learning methods, Support Vector Machine is superior to other machine learning methods. This is proven by the accuracy results obtained at 85% with feature extraction using ResNet-50 where the processing time is 8 minutes. Followed by the second-best model, namely ResNet-50 with XGBoost which obtained accuracy results of 82% with a processing time of 55 minutes. Meanwhile, the use of feature extraction using the DenseNet-121 method was obtained using a combination of the Support Vector Machine method and the XGBoost method with the accuracy obtained being 81%
AN IOT-BASED AUTOMATED WATERING SYSTEM FOR PLANTS USING INTEGRATED FUZZY LOGIC AND TELEGRAM BOT
Full text linkThe development of automatic plant watering systems has recently gained popularity due to the need to conserve water and ensure healthy plant growth. This study focuses on integrating fuzzy logic, sensors, and algorithms to provide an automatic watering system. Fuzzy logic is a powerful tool that allows the system to interpret sensor data and make informed decisions. The sensors measure soil moisture, humidity, temperature, and light intensity. The data collected from these sensors is analyzed using algorithms to determine the appropriate watering schedule. The system’s ability to analyze and interpret data ensures that the plants receive the necessary moisture without over-watering or under-watering. Integrating the Telegram Bot is a significant feature of the system, enabling users to monitor and control the system remotely. The Telegram Bot sends users notifications when the system is activated, or the plants require attention. The system can also be controlled remotely through the Bot, enabling users to adjust the watering schedule or turn the system on or off. This research shows that the designed features of the system function effectively and can be used on a daily household scale. The system’s automated features reduce the need for constant monitoring and manual watering, making it ideal for those who engage in gardening at home. This innovation is particularly relevant in increasing the productivity of plants. In addition, the system’s ability to be controlled remotely through the Telegram Bot is a significant advantage, making it accessible and convenient for users
SOFTWARE DEFECT PREDICTION USING PCA BASED RECURRENT NEURAL NETWORK
Full text linkSoftware quality is one of the important phases in software development. Software quality assesses the usability and quality of the software developed. Defect prediction early in software development helps in software quality assurance by reducing software defects that may occur. With good predictions, it will provide additional benefits in terms of resource and cost efficiency. The researchers in this study have proposed a software defect prediction method that utilizes a Recurrent Neural Network (RNN) based on Principal Component Analysis (PCA). The dataset used is the PROMISE dataset, namely JM1, CM1, PC1, KC1, and KC2. The test results showed that the PCA-RNN method was successfully applied. For the highest accuracy on the PC1 dataset, with an accuracy of 93.99% with the division of training data by testing data (70:30)
OPTIMIZING SENTIMENT ANALYSIS IN EDUCATIONAL YOUTUBE VIDEOS: A COMPARATIVE STUDY OF ROBERTA AND MULTINOMIAL NAIVE BAYES
Full text linkYouTube has evolved into a globally influential platform, engaging over 2.1 billion users worldwide and serving as a prominent medium for sharing, consuming, and creating diverse video content. Particularly popular among younger demographics, YouTube stands as a multifaceted hub spanning various genres and has significantly impacted education by providing extensive educational materials, fostering independent learning, and supporting a wealth of educational resources. This research conducts an in-depth investigation into sentiment analysis specifically within the context of educational YouTube videos. Leveraging advanced machine learning techniques, notably RoBERTa, this research conducts a comparative analysis with Multinomial Naive Bayes (MNB). The primary focus is on exploring RoBERTa\u27s adaptability and performance across a spectrum of educational video content, revealing its commendable accuracy of 91.21%, surpassing MNB\u27s accuracy of 79.59%. However, it is observed that RoBERTa\u27s performance is notably affected by smaller datasets, highlighting the critical importance of ample and diverse training data for achieving optimal results. These findings highlight the pivotal role of dataset characteristics and size in developing robust sentiment analysis models, especially with advanced natural language processing methods like RoBERTa