Risqy Pradana Putra, Fara Triadi -, Ahmad Rofiq Hakim
Perkembangan teknologi informasi mendorong penerapan sistem yang lebih efisien dan transparan dalam manajemen gudang. Penelitian ini bertujuan merancang sistem manajemen gudang berbasis QR Code dan Blockchain untuk meningkatkan akurasi, keamanan, dan efisiensi pelacakan barang. Sistem mengintegrasikan mikrokontroler ESP32, modul GM65 barcode scanner, printer thermal, dan UPS sebagai sumber daya mandiri. QR Code digunakan untuk identifikasi dan pelacakan barang secara real-time, sedangkan Blockchain memastikan data transaksi tersimpan secara aman, transparan, dan tidak dapat diubah. Penelitian menggunakan metode Waterfall yang meliputi analisis kebutuhan, perancangan, implementasi, dan pengujian sistem. Hasil pengujian menunjukkan bahwa sistem mampu melakukan pencatatan, pemindaian, dan pembaruan data stok secara real-time dengan tingkat akurasi yang tinggi. Sistem ini memberikan solusi yang efektif untuk meningkatkan efisiensi operasional, keamanan data, dan transparansi dalam pengelolaan gudang berbasis Internet of Things (IoT).
Ryan Anthony, Jechenthia Maria Taso, Stephen Yohanes Christopher, Ridhwan Ardiyansyah
This study aims to analyze and compare the performance of three algorithms, namely Support Vector Regression (SVR) with a linear kernel, XGBoost, and LightGBM, in predicting the Price of Ethereum cryptocurrency based on daily historical data. The study uses Ethereum Price data in USD for the last five years obtained from the investing.com website. The variables used are Close, Open, High, and Low Prices. The study uses two data splitting scenarios: 80% training data and 20% testing data, and 70% training data and 30% testing data. This study also uses time step variations to test the effect of time dependency on algorithm performance. The results indicate that the LightGBM algorithm has the best performance compared to the other two algorithms with an average MAE value for High Price of 75.486, SVR has a value of 115.590, and XGBoost has a value of 77.314 in the 80% training data and 20% testing data split. In the 70% training data and 30% testing data split, the LightGBM algorithm still excels with an average MAE value for High Price of 78.228, SVR of 104.356, and XGBoost of 83.573. Other evaluations such as RMSE and R2 also show the superiority of the LightGBM algorithm. For the required computation time, the SVR algorithm outperforms the other two algorithms.