Publication Search

This research aims to develop a multiclass flower image classification system using the Convolutional Neural Network (CNN) algorithm with the EfficientNet architecture. The main problem addressed is the difficulty of manual identification of flower species that share high visual similarity. The research stages include collecting 17,299 flower images across 19 classes, performing data preprocessing such as image resizing, pixel normalization, and augmentation, followed by model training using the EfficientNet transfer learning approach. The model was trained for 10 epochs with an 80:20 training-validation data split. The evaluation results show that the model achieved a validation accuracy of 98.05% with a loss value of 0.0968, and an average precision, recall, and F1-score of 0.98. The trained model was then implemented into a web-based application built using the Next.js framework, enabling users to upload flower images and obtain real-time classification results via the Hugging Face API. The system successfully identified flower species with a confidence level of 99.87%. These findings demonstrate that combining a modern CNN architecture with transfer learning provides efficient and highly accurate flower classification performance, which can be effectively implemented for educational and digital conservation purposes.

Detection and identification of plant diseases is critical to the success and efficiency of agricultural production. Plant disease outbreaks are becoming more frequent throughout the world, and the presence of these diseases in cultivated plants has a significant impact on productivity. Therefore, researchers are focusing on developing effective and reliable plant disease detection methods. Thus, farmers can take advantage of early detection of this disease to minimize future losses. This article discusses machine learning approaches as well as decision trees, K-nearest neighbors, naive Bayes, support vector machines (SVM), and random forests for detecting coffee leaf diseases using leaf images. The above-mentioned classifications were researched and compared to determine the most suitable plant disease prediction model with the highest accuracy. Compared with other classification algorithms, the SVM algorithm achieves the highest accuracy of 99.75%. All the models trained above will be used by farmers to quickly identify and classify new diseases in images as a prevention strategy. As a preventive measure, farmers can detect and classify new diseases in images early.

The performance of the Indonesian National Team (Timnas) in the 2026 World Cup qualifications has triggered massive and diverse responses on social media, particularly on platform X. This study aims to identify and classify public sentiment regarding Timnas Indonesia's performance into positive, negative, and neutral categories using a data mining approach. Text data was processed through pre-processing stages, term weighting using TF-IDF, and the application of the Synthetic Minority Over-sampling Technique (SMOTE) to address significant class distribution imbalance. The classification algorithm employed was Multinomial Naïve Bayes. Model performance evaluation was conducted by comparing two training-testing data split scenarios: 90:10 and 80:20 ratios. The results indicate that public opinion is dominated by negative sentiment at 73.2%, reflecting public disappointment. In terms of model performance, the 90:10 ratio scenario yielded the best accuracy of 80%, outperforming the 80:20 ratio which recorded an accuracy of 75%. These findings demonstrate that combining Multinomial Naïve Bayes with the SMOTE technique is effective in handling imbalanced text data and is capable of accurately mapping public perception.

Manual quality assessment of Platelet Concentrate (TC) is highly subjective and inconsistent, necessitating an objective, automated classification system. This study aims to develop a computationally efficient, low-cost model for TC quality classification using Histogram Features extracted from grayscale images combined with the K-Nearest Neighbor (KNN) algorithm. The methodology employed critical preprocessing steps, including StandardScaler for normalization and SMOTE for balancing the training data, followed by optimization across K=1 to K=30. The optimal model achieved a maximum accuracy of 69.23% at K=6, with an F1-Score of 71.43%, confirming robust performance on the imbalanced testing set. The results validate the effectiveness of the Histogram-KNN approach as a consistent and reliable decision support system for rapid TC quality screening in resource-limited settings.

Tourism is one of the sectors that plays an important role in boosting economic growth through travel activities and destination exploration. Tourists' preferences for nature-based tourism options, such as mountain hiking or beach tourism, are influenced by various factors, ranging from personal experiences and recreational interests to social characteristics. Therefore, a technology-based approach is needed to predict destination choice tendencies more accurately. As artificial intelligence technology develops, deep learning methods have been widely used in classification processes due to their ability to process large amounts of data and recognize complex patterns. In this study, a Multilayer Perceptron (MLP) model is used to classify tourists' preferences between mountain or beach destinations based on a survey dataset. The research stages include data processing, data splitting using a train-test split, model training, and performance evaluation using accuracy, precision, recall, and F1-score. The test results show that the MLP model is capable of achieving an accuracy rate of 99%, confirming that deep learning methods are effective in automatically mapping tourism preference trends. This research is expected to serve as a basis for the development of more personalized travel destination recommendation systems, as well as to support tourism management in formulating targeted promotional strategies.

Relational databases store information in interconnected tables and are widely used for data management and retrieval. However, in certain environments, the original values stored in a relational database cannot be exposed during data retrieval. This limitation creates a challenge because common encryption methods only transform data for storage and do not support mathematical operations needed for value matching. Partially Homomorphic Encryption is a cryptographic approach that allows specific mathematical operations to be performed directly on transformed data without restoring it to its original form. This study proposes the use of Partially Homomorphic Encryption to enable value-based data retrieval while keeping all stored values in their transformed form throughout the entire process. The method relies on homomorphic properties that allow mathematical comparison to be conducted on encrypted data, making the retrieval process possible without revealing the original values. The results show that this approach can perform data retrieval operations in a relational database while preserving the transformed structure of the stored data. The proposed method offers an alternative for environments that require data retrieval without exposing original values and demonstrates the potential of homomorphic techniques in supporting secure and functional data processing in relational database contexts.

Indonesian Sign Language (BISINDO) serves as a primary communication medium for the deaf community; however, limited public understanding often creates barriers during daily interactions. This study aims to develop a real-time BISINDO word-level translation system using hand landmark extraction and temporal modeling with Long Short-Term Memory (LSTM). The system employs MediaPipe Hands to detect 21 hand landmarks per frame, which are then processed as sequential motion patterns to classify five BISINDO words: saya, terima kasih, maaf, nama, and kamu. A total of 250 gesture samples were recorded under controlled lighting conditions as the primary dataset. The processed sequences were used to train the LSTM model, which was subsequently integrated with an ESP32 microcontroller and a DFPlayer Mini module to produce direct audio output. Experimental results show that the model achieved an average accuracy of 86%, with precision and recall values ranging from 0.81 to 0.94. The confusion matrix analysis indicates that most gestures were correctly classified, although some errors occurred in gestures with similar initial motion trajectories. Integration testing demonstrated an average system latency of 3.8 seconds and an audio output success rate of 85%. These findings indicate that the proposed system is capable of translating BISINDO word-level gestures accurately, responsively, and consistently in real-time conditions. This study provides a strong foundation for the broader development of sign language translation systems, with potential enhancements in vocabulary expansion, multi-user datasets, and hardware optimization for deployment in real-world environments.

Indonesia has extraordinary resources and potential in developing renewable energy sources (RES), but various obstacles must be overcome in implementing RES. The purpose of this study is to analyze the gap in the application of RES. This gap includes energy knowledge, community participation, battery waste management, service quality, regulation, and legal policy. This study uses a mixed-methods approach, by conducting a structured questionnaire in quantitative data collection, while qualitative data collection through special interviews, focused group discussions, and conducting policy regulation analysis. The results show that 62% of people do not understand RES, 28% are involved in project planning, and 74% are unaware of SOP (standard operating procedures) regarding battery waste recycling. The results of the correlation analysis reveal a positive relationship between the level of knowledge and interest in RES (R = 0.56). Also, the developed community-based participation model includes initial involvement, transparency of information, and local incentives. These findings further strengthen the compatibility of the innovation diffusion theory, planned behavior theory, SERVQUAL, and the theory of public interest. This study will make a practical contribution through evidence-based strategies in increasing resilience, especially for policymakers and energy service providers. The impact of the policy aspects includes the need for large reforms, education, public campaigns, and the realization of battery waste management systems. This study also provides an opportunity for further study by expanding the geographical scope and related industrial sectors.

Improving the accuracy of emission monitoring in the Continuous Emission Monitoring System (CEMS) is crucial to support compliance with environmental regulations, especially in gas and steam-based power plants (PLTGU). At PT Jawa Satu Power's PLTGU, the purging system on the CEMS sample probe is still performed manually and limited to scheduled preventive maintenance, potentially reducing emission data reliability due to particle contamination. This study aims to design and simulate an automatic purging system based on a Programmable Logic Controller (PLC), taking into account technical parameters such as timer, gas pressure, temperature, and sample flow rate. The system design was carried out through the modeling of automatic control logic using CX-Programmer software, with a protection approach based on real-time conditions and timing. The design results show that the automatic purging system can improve cleaning consistency, reduce the risk of contamination, and enhance the integrity of emission monitoring data. This study is expected to serve as a foundation for developing a more applicable and integrated automated purging system for CEMS in the future.

This study presents a machine learning approach for predicting the dimensions of microstrip antenna slots based on antenna performance parameters such as frequency, gain, directivity, return loss (S11), radiation efficiency, and VSWR. A two-phase methodology was employed. In the first phase, ten regression algorithms were evaluated, and Random Forest was identified as the most effective model based on Mean Absolute Error (MAE) and R-squared (R²) scores. In the second phase, hyperparameter tuning was conducted using Grid Search to further improve the model’s performance. The optimized Random Forest model demonstrated consistent improvements in predictive accuracy, with R² values increasing across all output variables. These results indicate that the combination of regression-based modeling and systematic hyperparameter tuning is effective for capturing complex relationships in antenna design tasks. The proposed approach offers a promising data-driven alternative for geometric prediction in microstrip antenna development, particularly when analytical models are insufficient.

Terrestrial digital television transmitter stations are the main facilities in the transmission of digital television broadcasts to the public. The quality of the transmitted signal is expected to reach the Central Java-1 service area well so as to provide optimal and reliable quality of digital television broadcast performance according to the needs of the community, but currently, complaints about signal problems such as service coverage and reception quality still occur a lot, coverage and signal quality received by community-owned television transmitters cannot be separated from the influence of the quality performance of digital television transmission stations. The purpose of this research is to analyse the coverage performance of digital television services of digital television transmitter stations using the K-Means Clustering Method, identify areas with the best signal coverage and group areas based on the level of signal performance. The data used includes field strength parameters collected through field measurements at 25 service area location points, topography factors and transmitter distance were found to be the main causes of signal quality differences. Data analysis shows that the K-Means Clustering method effectively clusters areas with signal reception quality categories of very good cluster 3 areas, good cluster 8 areas, fair cluster 5 areas and poor cluster 9 areas. The results of this study can be recommended in the evaluation and optimisation of tele-transmitting station networks.

Peningkatan transaksi dalam platform e-commerce seperti Shopee menuntut adanya sistem prediksi status pesanan yang akurat, guna mengoptimalkan pelayanan dan mengurangi pembatalan maupun keterlambatan pengiriman. Penelitian ini bertujuan membangun model klasifikasi status pesanan (selesai atau batal) pada toko Stuftech.Id menggunakan algoritma C4.5. Data yang digunakan merupakan transaksi pesanan mencakup metode pembayaran, kategori wilayah pengiriman, dan ongkos kirim. Proses klasifikasi dilakukan menggunakan RapidMiner dengan tahapan preprocessing, pembangunan decision tree, dan evaluasi model. Hasil analisis menunjukkan bahwa atribut “Kategori Pulau” memiliki nilai gain tertinggi sehingga dipilih sebagai node akar. Model yang dibentuk menghasilkan akurasi sebesar 86%, dengan recall 100% untuk pesanan selesai namun hanya 6,67% untuk pesanan batal. Temuan ini mengindikasikan bahwa algoritma C4.5 efektif dalam memprediksi pesanan yang berhasil, namun perlu peningkatan dalam mendeteksi potensi pembatalan. Implementasi model ini dapat membantu pelaku usaha dalam mengambil keputusan operasional secara proaktif.

The coffee farming sector in Gunungmanik Village, Indonesia, plays a significant role in the local economy. However, the monitoring and management of coffee crops remain largely manual and conventional, making it difficult for farmers to respond quickly to environmental threats such as drought, pests, or sudden temperature shifts. This research presents the development of iotgm.id, a web-based monitoring system integrated with Internet of Things (IoT) devices designed to provide real-time environmental data for coffee plantations. The system measures key parameters including temperature, soil moisture, and motion detection (as a proxy for pest activity), and delivers this data via a user-friendly web interface. It also features digital farm record management, real-time alerts for abnormal conditions, and data visualization through interactive dashboards. Field testing with local farmers showed that the system improves decision-making, speeds up responses to environmental changes, and reduces the need for direct field visits. Unlike earlier systems that often required technical expertise or focused on single parameters, this system offers multi-parameter monitoring and is accessible to farmers without advanced digital literacy. The system bridges the gap between sophisticated agricultural technologies and practical field-level application. It contributes to the adoption of precision agriculture in rural areas, offering a scalable model for broader implementation in similar contexts

The insurance business in an insurance company offers insurance products owned by insurance companies. There are many transactions such as the purchase of insurance products and the application of disbursement of insurance benefits to customers, so that disputes occur in the reports generated every month and this does not become effective and time-efficient as needed. This study aims to evaluate the effectiveness of a website-based real-time insurance transaction reporting monitoring system. This research method involves the development of a web-based system designed to monitor and report insurance transactions directly, as well as the evaluation of system performance using quantitative and qualitative approaches. The research stages include needs analysis, system design and development, implementation, and system testing and evaluation in insurance companies. The results of the study show that a website-based system can facilitate evaluation Monitoring the results of reports on ongoing transactions, so that reports every month can be formed digitally through the system that has been created.  

Kecelakaan lalu lintas di Kabupaten Brebes merupakan masalah kritis karena tingginya frekuensi insiden yang terjadi di wilayah tersebut. Penelitian ini bertujuan untuk menentukan area yang rentan terhadap kecelakaan dengan menggunakan algoritma K-Means Clustering , yang mendukung proses pengambilan keputusan berbasis data. Isu utama yang dieksplorasi dalam penelitian ini adalah bagaimana algoritma K-Means dapat diimplementasikan untuk mengelompokkan zona rawan kecelakaan dan meningkatkan kesadaran masyarakat terhadap keselamatan jalan. Metodologi yang digunakan meliputi pengumpulan data melalui tinjauan pustaka, observasi langsung, dan wawancara, yang dilanjutkan dengan penggunaan algoritma K-Means untuk mengklasifikasikan data kecelakaan berdasarkan jumlah kejadian, korban jiwa, dan cedera. Temuan menunjukkan bahwa algoritma K-Means secara efektif mengelompokkan lokasi rawan kecelakaan ke dalam tiga tingkat risiko yang berbeda: tinggi, sedang, dan rendah. Dengan demikian, informasi yang terklasifikasi ini dapat membantu otoritas terkait dalam meningkatkan langkah-langkah keselamatan lalu lintas dan mengedukasi masyarakat tentang area berisiko tinggi. Hasil penelitian ini diharapkan dapat berkontribusi pada pengembangan kebijakan keselamatan lalu lintas yang lebih terinformasi dan strategis di Kabupaten Brebes.

The election of the OSIS chairman at SMK Al-Ikhlas Losari often influences subjectivity, so a goal-oriented system is needed to select the best candidate. This study applies the WASPAS method in a Decision Support System to broadcast OSIS chairman candidates based on criteria such as attendance, integrity, academic achievement, and health. Data were collected through observation, interviews, and literature studies, then processed using Microsoft Excel. The results of the WASPAS calculation yield a Qi preference value, which determines the ranking of candidates transparently. This system reduces subjectivity, increases efficiency, and ensures the selection of competent leaders, supporting the OSIS vision.

The advancement of information technology has had a profound impact on education, including at the Universitas Bina Sarana Informatika (UBSI), where My Best, an elearning application, was created. With this program, users may participate in a variety of academic tasks, including online attendance, discussions, assessments, assignments, and course materials. But pupils continue to face usability difficulties. The System Usability Scale (SUS) technique is used in this study to assess the usability of the My Best program. The approach included 35 current UBSI students who have been using the My Best software for more than seven semesters. According to the assessment, the average SUS score was 70, which is considered to be rather high but still needs work in terms of interface usability and user-friendliness. As a result, although the My Best app is thought to be helpful in fostering learning, improvements are needed to enhance the whole user experience.

Technological advancements have brought fundamental changes in the way we interact with digital images and photography. One significant milestone in this development is the Photoshop Express Photo Editor, which has become a primary platform for image processing and editing. Datasets are used to analyze sentiment and are utilized during the accuracy testing phase. Based on the testing results, the Convolutional Neural Network (CNN) algorithm achieved an average accuracy value of 86.50%, compared to the Naïve Bayes (NB) algorithm, which achieved an average accuracy value of 75%. The results of the research conclude that the choice of sentiment analysis method should be tailored to the needs and limitations of the system. If a fast, light, and easy-to-understand process is required, the Naive Bayes method is the right choice. However, if accuracy and context understanding are the top priorities, then CNN is a superior approach, although it requires more resources. Additionally, based on the Wordcloud data, it is known that the majority of comments are positive, indicating that the reviews or texts analyzed contain many positive expressions related to quality, usability, and ease of use.

This study focuses on developing a web-based Research and Community Service Information System (SIPPMAS) for Politeknik Takumi Bekasi, utilizing the Waterfall methodology. The aim is to create an integrated platform that streamlines the management of research and community service activities, from proposal submission and budget allocation to project execution and final reporting. The Waterfall method was chosen for its structured, sequential approach, ensuring a systematic development process through distinct phases: requirements analysis, design, implementation, testing, and maintenance. This approach is expected to enhance data accuracy, improve operational efficiency, and provide real-time project monitoring, ultimately facilitating better collaboration among stakeholders and increasing the overall impact of research and community service initiatives at Politeknik Takumi Bekasi. The system is designed to address current manual administrative challenges, offering a centralized and accessible solution for all users.

The digital transformation in education has encouraged the adoption of computer-based tests (CBT) using video content, which demands stable and efficient network performance. This study aims to evaluate the performance of two queue management algorithms, namely Random Early Detection (RED) and Per Connection Queue (PCQ), in maintaining the quality of service (QoS) of school networks during online video-based examinations. A case study approach was applied using a real network topology in a school environment, and QoS parameters such as throughput, delay, packet loss, and jitter were measured. The implementation was conducted using a MikroTik RB450Gx4 router configured with simple queue settings for each algorithm. The results show that PCQ provides more consistent performance under high user loads, achieving an average throughput of 56,482 bps and lower delay compared to RED. Conversely, RED performs better in scenarios with a small number of users. The study recommends using PCQ for networks with dynamic and dense user environments, while RED is more suitable for low-traffic conditions where latency stability is crucial. These findings offer practical guidance for managing bandwidth and improving the quality of CBT delivery in educational settings.