Jurnal Informatika: Jurnal Pengembangan IT
Not a member yet
437 research outputs found
Sort by
Adopsi dan Kehadiran Media Sosial Untuk Penanggulangan Bencana (Studi Pada Badan Penanggulangan Bencana Daerah (BPBD) Tingkat Provinsi di Indonesia)
No full textOrganisasi perlu mengadopsi dan aktif menggunakan media sosial agar dapat memanfaatkan sumber daya dengan masyarakat melalui media sosial sehingga dapat meningkatkan kinerja dalam mencapai tujuan organisasi. Walaupun studi mengenai partisipasi aktif masyarakat melalui media sosial dalam kegiatan penanggulangan bencana oleh lembaga kebencanaan tingkat nasional di Indonesia telah dilakukan, namun belum ada pemahaman menyeluruh mengenai tingkat adopsi dan penggunakan media sosial oleh lembaga kebencanaan tingkat provinsi. Untuk mengisi celah literature tersebut, kami menganalisis tahapan adopsi dan kehadiran di media sosial BPBD tingkat provinsi di seluruh Indonesia. Hasil analisis kami terhadap website dan akun media sosial 33 BPBD tingkat provinsi di Indonesia menunjukkan tingkat institusionalisasi media sosial yang cukup rendah dan adanya tingkat variasi kehadiran di media sosial. Kami menyajikan data terkini mengenai adopsi dan kehadiran BPBD tingkat provinsi di media sosial yang berguna bagi BNPB, BPBD dan pemerintah daerah untuk meningkatkan kinerja organisasi dalam penanggulangan bencana
Pemodelan Topik Komentar Terhadap Aplikasi Allstat BPS Tahun 2017-2025
No full textPenelitian ini dilatarbelakangi oleh meningkatnya kebutuhan akan data statistik yang mudah diakses melalui aplikasi mobile, salah satunya adalah aplikasi AllStat BPS. Tujuan dari penelitian ini adalah untuk menganalisis sentimen dan mengidentifikasi topik utama dalam ulasan pengguna aplikasi AllStat BPS pada periode 2017–2025. Metode yang digunakan mencakup analisis sentimen berbasis lexicon dengan kamus InSet dan klasifikasi menggunakan algoritma Naive Bayes, Random Forest, dan Support Vector Machine (SVM). Pemodelan topik dilakukan dengan pendekatan Latent Dirichlet Allocation (LDA). Hasil penelitian menunjukkan bahwa model Random Forest memberikan performa klasifikasi terbaik dengan akurasi pada data latih sebesar 88,16% dan nilai kappa 0,8046. Selain itu, LDA berhasil mengidentifikasi delapan topik utama dari ulasan pengguna, dengan Topik 1 memiliki nilai koherensi tertinggi (0,1784) yang mengindikasikan kekuatan semantik antar kata dalam topik tersebut. Topik-topik ini kemudian dipetakan ke dalam kerangka kualitas perangkat lunak berdasarkan standar ISO/IEC 25010, dengan aspek Functional Suitability dan Performance Efficiency sebagai topik dominan. Kesimpulan dari penelitian ini adalah bahwa kombinasi metode Random Forest dan LDA efektif dalam mengklasifikasikan sentimen serta menggambarkan fokus isu dalam ulasan pengguna aplikasi AllStat BPS
Prediksi Hasil Panen Karet di Gunung Tua Menggunakan Support Vector Machine
No full textPenelitian ini bertujuan untuk memprediksi hasil panen karet di wilayah Gunung Tua, Kabupaten Padang Lawas Utara, dengan menggunakan algoritma Support Vector Machine (SVM). Produksi karet dipengaruhi oleh berbagai faktor musiman dan kondisi lingkungan yang menyebabkan fluktuasi hasil panen, sehingga menyulitkan perencanaan bagi petani maupun instansi terkait. Penelitian ini menerapkan pendekatan supervised learning dengan metode Support Vector Regression (SVR) untuk memodelkan prediksi hasil panen karet berdasarkan data produksi historis yang diperoleh dari instansi pertanian setempat. Tahapan penelitian meliputi pengumpulan data, prapemrosesan, normalisasi data, pelatihan model, dan pengujian. Evaluasi kinerja model dilakukan menggunakan Root Mean Square Error (RMSE) sebagai indikator tingkat kesalahan prediksi. Hasil penelitian menunjukkan bahwa model SVM mampu memprediksi hasil panen karet dengan nilai RMSE sebesar 191 dan tingkat akurasi sebesar 96,2%, yang menunjukkan bahwa model memiliki performa yang baik dalam menangkap pola data produksi. Dengan demikian, algoritma Support Vector Machine dapat dimanfaatkan sebagai alat pendukung pengambilan keputusan dalam perencanaan dan pengelolaan produksi pertanian kare
Analisis Emosi Komentar Pengguna TikTok terhadap Film Jumbo Menggunakan Metode Naive Bayes
No full textTikTok has become a widely used social media platform where users actively express opinions through comment features. This study aims to classify the emotions contained in TikTok user comments on the Indonesian animated film Jumbo using the Naive Bayes Classifier method. The dataset consisted of 1,341 comments collected from the official Visinema Pictures account using the Apify Web Scraper. The collected data were processed through several preprocessing stages, including case folding, tokenization, normalization, stopword removal, and stemming using the Sastrawi library. Emotion labeling was performed based on the Indonesian NRC EmoLex lexicon by categorizing comments into three emotional classes: angry, happy, and sad. Feature extraction was conducted using the TF-IDF weighting method to generate relevant text representations and identify dominant terms in each emotional category. The dataset was divided into 80% training data and 20% testing data to evaluate the model performance. The experimental results show that the Naive Bayes model achieved an accuracy of 78.81%. The emotion distribution indicates that anger was the most dominant class with 904 comments, followed by happy with 415 comments, and sad with 22 comments. The model demonstrated the best performance in the anger class, achieving 100% recall, 75% precision, and an F1-score of 85.71%. However, the classification performance for minority classes, particularly happy and sad, still requires improvement. This research contributes to the development of text mining-based emotion analysis and provides insights into audience emotional responses that may support film evaluation and marketing strategies
Sistem Informasi Identifikasi Faktor yang Mempengaruhi Prestasi Siswa MAN Karo Menggunakan Pendekatan Data Mining
No full textStudent academic achievement is a key indicator of learning success and is influenced by various internal factors (such as motivation, learning style, and attendance) as well as external factors (such as parental support and learning facilities). This study aims to identify the main factors that determine students academic achievement by implementing a web-based information system using the Apriori algorithm. The study applied a quantitative descriptive approach using data from 300 students of MAN Karo, collected through questionnaires and academic records. The Apriori algorithm was implemented with a minimum support threshold of 0.05 and a minimum confidence threshold of 0.6, producing 1,621 association rules. The results indicate that a combination of internal and external factors is strongly associated with high academic achievement. The strongest association rule shows that students who attend tutoring classes tend to achieve high report card scores (?85) and have complete learning facilities, with a confidence value of 67.6% and a lift ratio of 3.12. Overall, internal factors such as motivation (77.33%) and high attendance (63.67%), along with external factors including parental support, learning facilities, and tutoring participation, significantly contribute to improving students academic performance. These findings demonstrate that the Apriori algorithm can effectively support schools in identifying dominant academic factors and can be used as a basis for data-driven decision making
Smart Finance: Desain dan Implementasi Sistem Keuangan Cerdas Real-Time Berbasis IoT untuk UMKM
No full textThe development of Internet of Things (IoT) technology and real-time data analytics provides opportunities to improve financial management efficiency for Micro, Small, and Medium Enterprises (MSMEs). However, most MSMEs in Indonesia still rely on manual bookkeeping, which is inefficient, prone to errors, and limits access to formal financing. This study aims to design and implement Smart Finance, an IoT-based intelligent financial system capable of processing transaction data automatically in real time. The research method includes system requirement identification, system design and device integration, application implementation, and system performance testing. The system was developed as a web-based application integrated with IoT devices such as ESP32-CAM to support automatic transaction recording, cash flow visualization, and digital financial report generation. The testing results indicate that the system can automatically record transactions with good accuracy, provide real-time financial dashboards, and deliver transaction notifications, thereby helping MSME owners monitor their financial conditions more quickly and transparently. The main contribution of this study lies in integrating IoT devices with a web-based financial recording system that enables automatic and real-time transaction recording, an approach that is still rarely implemented in MSME financial management. Although challenges related to internet connection stability remain, the developed system demonstrates potential in improving efficiency, transparency, and the quality of financial decision-making among MSMEs. This study concludes that Smart Finance can serve as a practical and adaptive digital financial solution to support the sustainability and competitiveness of MSMEs in the digital era
Deep Learning-Based Sentiment Analysis of Islamic Boarding School Google Reviews Using IndoBERT Variants and XLM-RoBERTa
No full textOnline reviews on platforms like Google Maps have become a crucial data source for analyzing public opinion and consumer behavior, including in the context of selecting religious educational institutions, specifically pesantren (Islamic boarding schools). This study aims to perform sentiment analysis to measure public perception towards pesantren located across the island of Java. The data were collected via web scraping, yielding a total of 8,577 reviews, which subsequently underwent essential text preprocessing steps including cleansing, case folding, tokenization, stopword removal, and stemming. The prepared dataset was then partitioned using the Stratified Train-Test Split method into 70% for training and 30% for testing.The research evaluated the performance of three pre-trained language models IndoBERT Base, IndoROBERTa Small, and XLM-RoBERTa, which were fine-tuned using the Focal Loss function. The training strategy prioritized saving the best model based on the neutral F1-score.The final evaluation on the unseen test data demonstrated that the IndoBERT Base model significantly outperformed the others, achieving the highest overall accuracy of 0.92 (92%). This strong balance confirms the model's excellent generalization ability, indicating no significant overfitting and successful mitigation of classification bias. The findings validate IndoBERT Base as the optimal model for sentiment classification of pesantren reviews. Future research is recommended to shift focus toward building a larger, more diverse dataset to further enhance model generalizability
Optimasi Fuzzy Logic Menggunakan Genetic Algorithm (GA) dalam Menentukan Program Diet dan Bulking
No full textThe increasing demand for accurate and personalized diet and bulking programs highlights the need for a reliable decision support system (DSS). This study aims to develop a fuzzy logic–based DSS optimized with a Genetic Algorithm (GA) to recommend diet, bulking, or maintenance programs tailored to individual conditions. The methodology involved designing fuzzy sets, formulating IF–THEN rules, applying the Mamdani inference method, and optimizing fuzzy parameters using GA. Data were collected from 50 adult respondents, and the system was tested using 10 input scenarios validated by fitness experts. The results revealed that the fuzzy system without GA achieved only 38% agreement with expert recommendations, whereas GA optimization significantly improved accuracy to 82%. Furthermore, GA refined membership functions and eliminated irrelevant rules, producing a more streamlined yet precise system. The web-based interface facilitated user interaction and interpretation of results, ensuring practical usability. In conclusion, integrating fuzzy logic with GA enhanced the accuracy and adaptability of the system for determining diet and bulking programs, establishing it as a promising decision-making tool that can be further expanded with additional personalization variables in the future
Rancang Bangun Aplikasi Jatah Makan Karyawan Berbasis Android Dengan Qr Code (Studi Kasus: PT Wijaya Karya Beton Majalengka)
No full textPengembangan aplikasi jatah makan karyawan berbasis Android dengan QR Code di PT WIKA Beton bertujuan untuk meningkatkan efisiensi dan akurasi pendistribusian jatah makan. Sistem manual yang ada seringkali menimbulkan kesalahan data dan manipulasi, sehingga diperlukan mekanisme yang lebih efektif. Aplikasi ini dirancang untuk memonitor pendistribusian jatah makan, meningkatkan transparansi, dan memastikan setiap karyawan menerima haknya sesuai ketentuan. Metodologi pengembangan yang digunakan adalah Extreme Programming (XP), yang meliputi perencanaan, perancangan, pengkodean, dan pengujian. Pengumpulan data dilakukan melalui observasi dan wawancara untuk memahami sistem yang berjalan dan menganalisis kebutuhan pengguna. Sistem yang diusulkan menggunakan QR Code untuk identifikasi karyawan, yang dipindai saat pengambilan jatah makan. Data QR Code dibuat secara online dengan format id#nama. Aplikasi ini memiliki fitur seperti scanner dan riwayat pengambilan jatah makan. Hasil pengujian menunjukkan bahwa aplikasi ini efektif dalam membatasi kecurangan dan memastikan data yang akurat. Aplikasi ini memberikan kemudahan bagi karyawan dan staf kantin, serta menghilangkan kesalahan pencatatan manual. Penelitian lanjutan akan fokus pada peningkatan antarmuka, penambahan fitur senter pada scanner QR Code, dan penggunaan database online. Dengan demikian, aplikasi ini memberikan keuntungan signifikan dalam transparansi dan akurasi pendistribusian jatah makan di PT WIKA Beto
Performance Improvement of Machine Learning Algorithm using PCA on IoV Attack
No full textIn the transportation industry, the Internet of Vehicles (IoV) is an advancement of the Internet of Things (IoT), allowing automobiles to connect to networks to provide a range of features. This connectivity transforms traditional vehicles into intelligent systems, fostering innovations like autonomous driving and traffic optimization. However, this increased connectivity exposes IoV to cybersecurity threats, particularly because the networks utilized are often public and lack robust security measures. Cyberattacks targeting IoV can involve data packet modification, traffic flooding, or spoofing, potentially disabling critical vehicle components, compromising passenger safety, and increasing the risk of accidents. Consequently, accurate and efficient attack detection systems are essential to counter these threats and ensure IoV security. This study leverages the CICIoV2024 dataset and applies Principal Component Analysis (PCA) to enhance computational efficiency in detecting IoV attacks. The algorithms employed in this research include Random Forest, AdaBoost, Logistic Regression, and Deep Neural Networks. Experimental results demonstrate that implementing PCA significantly improves computational efficiency across all algorithms while maintaining consistent accuracy and F1-Score, highlighting its effectiveness in securing IoV systems.