Journal of Informatics And Telecommunication Engineering
Not a member yet
373 research outputs found
Sort by
Implementation Augmented Reality On Banjarnegara Culture As A Learning Media Using Marker-Based Tracking
Full text linkLocal Content is a subject that introduces the local culture of a region. At SD Negeri 1 Paseh, Banjarnegara, the teaching of Local Content faces challenges due to the lack of interactive learning media and conventional teaching methods that rely heavily on textbooks. This makes learning less engaging, lowers students' interest, and affects their understanding of Banjarnegara's culture. Thus, this study aims to develop a learning application based on Augmented Reality using the marker-based tracking method, which utilizes 2D images as markers to display 3D visuals. This approach helps enhance students' understanding of Banjarnegara's local culture in Local Content learning. The functional testing of the MULOK application was conducted using Blackbox Testing, which successfully evaluated its functionality. Distance and angle testing showed that markers were detectable within a range of 5-200 cm and at optimal angles between 20 and 160 degrees. Response time testing revealed an input-to-output response time of 1-2 seconds. Pretest and posttest evaluations were also applied to measure the application's effectiveness and to determine the improvement in students' understanding. The results showed an average score increase of 83.53%, demonstrating that Augmented Reality-based learning is more effective compared to conventional methods. Students do not need to visit the locations where culinary and handicraft products are made directly to observe and learn about them. By simply using the Augmented Reality Mulok application, students can effectively study Banjarnegara culture as part of Mulok learnin
Comparative Analysis of the Performance of Four Symmetric Algorithms on Digital File Security
Full text linkInformation security is crucial to prevent misuse that could harm others. Information can be accessed through various electronic devices such as mobile phones, computers, and tablets in the form of text, images, audio, and video, whether public or confidential. In the digital era, image files are highly susceptible to authenticity risks as they can be easily shared through various communication media. This facilitates unrestricted digital file exchange, raising concerns about authenticity and the risk of modifications before reaching the recipient. Therefore, digital file exchanges require a security system to ensure that transmitted data remains original and intact. Cryptography is a field of study that protects data security in communication. It consists of algorithms and keys, where algorithms perform encryption and decryption, while keys enhance security levels. This study examines image encryption by using different key lengths with the same image, as well as encrypting images of varying sizes using the same key length, employing AES, DES, 3DES, and RC6 algorithms. The results show that the DES algorithm is the fastest in encryption and decryption compared to the other three algorithms. DES is 13.3% faster than 3DES and 10.2% faster than RC6. Additionally, the key length used does not significantly impact processing time, but image size greatly affects encryption and decryption speed. These findings indicate that in cryptographic implementations for digital images, file size is a critical factor to consider to maintain efficiency without compromising encryption and decryption spee
Classification Of Outstanding Students Using Support Vector Machine (SVM) Based on Data Mining
Full text linkThis research aims to classify outstanding students at the Pagar Alam Institute of Technology using the Support Vector Machine (SVM) algorithm based on data mining. Early identification of outstanding students is crucial for supporting potential development and institutional decision-making. Historical data from 245 students from the 2016 to 2018 cohorts were utilized, encompassing course grades and Cumulative Grade Point Average (CGPA). The research process included data preprocessing such as normalization and splitting the data into 80% training data and 20% testing data. The SVM model was implemented with a Radial Basis Function (RBF) kernel and parameters C=1.0 and gamma=0.1. Evaluation results show that the model achieved an overall accuracy of 89.80% on the testing data. The model's performance was further validated through a confusion matrix (9 True Positives, 1 False Negative) and a classification report indicating good precision and recall for both classes. Furthermore, an Area Under the Curve (AUC) value of 0.93 signifies the model's excellent discriminative ability. This study contributes by providing an effective classification tool for identifying outstanding students, which can serve as a basis for the institution to design more targeted development and recognition programs
Comparison of Random Forest, K-Nearest Neighbors, Decision Tree, and Neural Network for Predicting Rainfall
Full text linkErratic rainfall due to climate change has significant impacts on the environment, agriculture and economy. To mitigate these impacts, a reliable rainfall prediction model is needed. Erratic rainfall due to climate change affects various sectors of life, so a reliable prediction model is needed to support data-based decision making. This study aims to compare the performance of Random Forest, k-Nearest Neighbors (kNN), Decision Tree, and Neural Network algorithms in predicting rainfall using observation data from the Citeko Meteorological Station. The data used include weather parameters such as temperature, humidity, and air pressure. The analysis was carried out using Orange software to evaluate the accuracy, precision, and computation time of each model. The results showed that Random Forest had the highest accuracy, while Neural Network showed consistent performance on more complex datasets. The kNN algorithm gave good results with the optimal number of neighbors, but was less efficient on large datasets. Decision Tree was easier to interpret but was prone to overfitting. This study provides insight into the most appropriate algorithm for rainfall prediction based on the characteristics of the data available
Sentiment Towards Social Media Politeness Ambassadors: A Case Study Using the Naive Bayes Method
Full text linkSocial media has had a significant impact on modern society, serving as a primary platform for sharing information and opinions. One intriguing phenomenon is the viral case of a female police officer, Putri Cikita, who earned the title "Ambassador of Courtesy" due to her actions in a video. This study aims to analyze public sentiment regarding this case on Twitter using the Naive Bayes Classifier (NBC) method. The research adopts a quantitative descriptive approach with sentiment analysis based on Text Mining, utilizing Python and Google Colab. The dataset consists of 2,000 Indonesian-language tweets collected from August to November 2024 using the keywords "Ambassador of Courtesy" and "Putri Cikita." The research stages include data collection, data preprocessing (case folding, tokenizing, filtering, stemming), and sentiment labeling into positive, negative, and neutral classes. The analysis results reveal that 11.55% of tweets express positive sentiment, 68.40% are neutral, and 20.05% are negative. The Naive Bayes method proves effective in classifying textual sentiment data. This research provides insights into public perceptions of viral events and underscores the importance of public image management in the digital era
Mobilenetv2 Analysis in Classification Diseases On Mango Leaves
No full textThis study aims to analyze the performance of the MobileNetV2 model in classifying diseases on mango leaves, consisting of three classes: capmodium, collectricu, and normal leaves. The dataset used contains 1500 images, with 80% allocated for training data, 10% for testing data, and 10% for validation data. The model was trained using a deep learning approach to identify mango leaf diseases based on the visual patterns present in each class. The results show that the MobileNetV2 model achieved an accuracy of 90%, a precision of 91%, a recall of 90%, and an F1-score of 89%. These findings highlight the potential of MobileNetV2 as an effective tool for automatically detecting mango leaf diseases. Therefore, this study is expected to contribute to the development of technology-based solutions in the agricultural sector, particularly in supporting farmers in identifying diseases quickly and accurately, thereby improving mango crop productivity
Application of MobileNetV2 Architecture with SIMAM for Automatic Detection of Diseases on Mango Leaves
Full text linkEarly detection of diseases in mango plants is crucial for improving crop yields and reducing economic losses for farmers. This study proposes the use of the MobileNetV2 architecture integrated with the Simple Attention Module (SIMAM) to enhance the accuracy of disease detection on mango leaves. MobileNetV2 was chosen for its computational efficiency, particularly on mobile devices, while SIMAM was utilized to strengthen the model’s focus on important visual features that represent disease symptoms on the leaves. The dataset used in this research consists of 3,000 images of mango leaves categorized into three classes: Capnodium, Colletotrichum, and Healthy Leaves. The model was evaluated using accuracy, precision, recall, and F1-score metrics. Experimental results show that the MobileNetV2 + SIMAM model achieved high performance, with an accuracy of 0.9833, precision of 0.9841, recall of 0.9833, and F1-score of 0.9833. With its combination of computational efficiency and high classification accuracy, this model has strong potential for implementation in mobile applications to assist farmers in detecting mango leaf diseases quickly, accurately, and practically in the field
Detection of Chicken Egg Quality with Digital Image using EfficientNet-B7
Full text linkChicken eggs are one of the staple food ingredients in Indonesia, playing a vital role in fulfilling the nutritional needs of the community. Therefore, an efficient, accurate, and reliable method for assessing egg quality is essential, especially to support the distribution process in the food industry. This study aims to develop a digital image-based classification system for assessing the quality of chicken eggs using deep learning methods with the EfficientNet-B7 architecture. EfficientNet-B7 was selected for its proven high accuracy in image classification tasks through the application of compound scaling, which simultaneously optimizes depth, width, and resolution. The dataset used in this study combines images collected from public sources and primary documentation, representing various conditions commonly found in chicken eggs. The preprocessing stage involved trimming techniques to focus on the egg object, followed by data augmentation using ImageDataGenerator, including rotation, shifting, zooming, and flipping to enhance dataset diversity. Model training was carried out with the early stopping technique to prevent overfitting. The experimental results showed that the model achieved an accuracy of 98.08% in classifying egg quality based on shell condition and other visual indicators. These findings demonstrate that the implementation of the EfficientNet-B7 model has great potential to support the automation of chicken egg quality assessment processes in a faster and more consistent manner. Thus, this research is expected to contribute to improving the efficiency of the food industry, particularly in the distribution process of chicken eggs in Indonesia
Analysis of Combined Contrast Limited Adaptive Histogram Equalization (CLAHE) and Median Filter Methods for Enhancement of CCTV Screenshot Image Quality
Full text linkThe quality of CCTV images often deteriorates due to poor lighting, low-quality cameras, and noise, hindering effective security analysis. This study aims to assess the combined effect of Contrast Limited Adaptive Histogram Equalization (CLAHE) and median filtering on improving the quality of CCTV screenshot images by enhancing contrast and reducing noise. Using a quantitative approach, four low-quality CCTV images were processed with CLAHE to improve contrast, followed by median filtering to reduce noise. Image quality was evaluated using two metrics: Mean Squared Error (MSE) and Peak Signal-to-Noise Ratio (PSNR). Results showed that CLAHE significantly improved image contrast, with MSE values ranging from 17.7513 to 159.092 and PSNR from 39.4809 to 47.1987. After applying the median filter, MSE values decreased to 12.1238–22.1747, and PSNR increased to 34.7288–37.3442, indicating noise reduction. The combination of CLAHE and median filter showed even better results, with MSE values ranging from 0.000993935 to 0.00508972, and PSNR ranging from 71.1032 to 78.1966. This combination significantly improved the quality of the CCTV screenshots, making them more suitable for security and forensic analysis. The findings suggest that CLAHE and median filtering can effectively enhance image clarity. Future studies should focus on optimizing these techniques for various lighting conditions and exploring other methods to address extreme noise levels in CCTV image
Coffee Quality Classification Based on Customer Reviews Using C4.5 Algorithm
Full text linkCoffee is a very popular commodity throughout the world, and its quality is oken evaluated through customer reviews. This research aims to classify coffee quality based on reviews given by consumers using the C4.5 algorithm. C4.5 is a machine learning algorithm used to generate decision trees, which allows decision making based on relevant attributes. In this research, the data used consists of customer reviews taken from e-commerce plaVorms and coffee discussion forums. The data is then processed with natural language processing (NLP) techniques to extract important features such as sentiment, keywords and term frequency. These features are used as input for the C4.5 algorithm, which builds a classification model based on patterns contained in the data. The results of the research show that the C4.5 model is able to classify coffee quality with high accuracy, reaching up to 85%. The factors that most influence quality classification include taste, aroma, and packaging, which are frequently mentioned in reviews. In addition, the analysis also shows significant differences in the quality of coffee produced from different coffee producing regions, which can provide insight for producers to improve their products