Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
Not a member yet
1071 research outputs found
Sort by
Sistem Pemantauan dan Pengendalian Logistik Buah Mangga Berbasiskan Machine Learning
No full textFruits are highly perishable goods, which means they have a short shelf life and can pose significant challenges in trade. A long supply chain can trigger the process of fruit spoilage. The logistics environment, both internal and external, can also affect the decrease in quality of goods. One common issue facing producers is the variability in consumer demand for fruit quality. To address this problem, a machine learning-based logistics monitoring and recommendation system can be developed, utilizing the Long Short-Term Memory (LSTM) and Decision Tree algorithms. Using machine learning algorithms, the system can analyze data from devices equipped with the Internet of Things (IoT), such as temperature and humidity sensors, to identify potential issues in the supply chain and provide recommendations to optimize logistics operations. In this study, a machine learning-based monitoring system is developed to monitor the shelf life of perishable goods, with a specific focus on mango fruit. The system utilizes LSTM to predict mango ripeness and decision tree algorithms to recommend fruit ripeness. The objective is to provide producers with recommendations that optimize the logistics process for high-quality mangoes and meet the consumer demands for quality fruit. The implementation of a machine learning-based logistics monitoring and recommendation system can provide significant benefits to mango producers. Using advanced technologies, such as LSTM and Decision Tree algorithms, producers can optimize their logistics operations, improve fruit quality, reduce waste, and improve customer satisfaction.Buah-buahan memiliki sifat yang mudah rusak dan dalam perdagangan memiliki permasalahan masa simpan yang relatif singkat. Rantai pasok yang cukup panjang dapat memicu proses kerusakan buah. Suhu dan waktu merupakan dua faktor yang dapat menentukan perubahan kualitas buah, oleh karena itu dapat digunakan suhu penyimpanan optimal selama proses logistik. Logistik yang baik adalah kondisi dimana kualitas barang dapat terjaga sampai tiba pada pelanggan. Lingkungan logistik, baik dari internal atau eksternal dapat mempengaruhi penurunan kualitas barang. Beberapa permasalahan yang sering dihadapi oleh rumah kemas di antaranya yaitu beragamnya kebutuhan konsumen terhadap kualitas buah. Maka itu, dibutuhkan sistem monitoring dan rekomendasi logistik berbasiskan machine learning dengan menggunakan algoritma Long Short-Term Memory (LSTM) dan Decision Tree untuk memberikan rekomendasi kepada produsen agar proses logistik kualitas buah mangga dapat memenuhi kebutuhan konsumen terhadap kualitas buah
A Comparative Study of HTTP and MQTT for IoT Applications in Hydroponics
Get PDFHydroponics is based on nutrients in water. It must be regularly monitored to prevent plant defects. The Internet of Things has become a solution for remote hydroponic monitoring and is currently being tested on the Yuan Hidroponik Kelompok Wanita Tani (KWT). This system will send data every minute, and each data has a possibility of loss in transmission. There is a chance that this system will be implemented in other hydroponic organizations. As more devices are involved, it will affect server resources. This research will compare Message Queue Telemetry Transport (MQTT) and Hypertext Transfer Protocol (HTTP) as popular protocols used in IoT. A test with increasing clients shows that at 50 clients HTTP needs 87% CPU, while MQTT needs 22.63% CPU. A test with increasing payload shows that at 10,000 payload HTTP needs 94% CPU while MQTT needs 28.35% CPU. A test with fixed clients and payloads shows that HTTP has a CPU limit based on the clients involved. A transfer time test shows that HTTP needs 177.344 seconds while MQTT needs 3.24 seconds. An acceptance rate is calculated by incrementing the count for every incoming payload. It shows that HTTP can receive 30,000 payloads, unlike MQTT which can only receive 1680 payloads before losses
Making AI Work for Government: Critical Success Factor Analysis Using R-SWARA
No full textThis study quantifies what makes Artificial Intelligence (AI) work for government, the critical success factors (CSFs) for successful AI implementation within the Directorate General of Taxes (DGT). Analyzing factors such as technology, organization, process, and environment, the research highlights the importance of organizational readiness, strategic vision, and leadership support to drive successful AI integration within DGT. The dimension of the organization became the most critical factor, followed by technology, process, and environment. The findings offer actionable insights for DGT's decision-making processes, aiding in strategic resource allocation and tailored AI strategy refinement. Furthermore, this research is a valuable reference for other public sector organizations that aim to enhance operational efficiency through the adoption of AI. This study empowers decision makers within the DGT and the wider public sector by providing nuanced information on the critical factors that influence the successful implementation of AI, fostering improved operational efficiency and governance practices
Comparative Analysis of Gradient Descent Learning Algorithms in Artificial Neural Networks for Forecasting Indonesian Rice Prices
No full textArtificial Neural Networks (ANN) are a field of computer science that mimics the way the human brain processes data. ANNs can be used to classify, estimate, predict, or simulate new data from similar sources. The commonly used algorithm for prediction in ANN is Backpropagation, which yields high accuracy but tends to be slow during the training process and is prone to local minima. To address these issues, appropriate parameters are needed in the Backpropagation training process, such as an optimal learning function. The aim of this study is to evaluate and compare various learning functions within the Backpropagation algorithm to determine the best one for prediction cases. The learning functions evaluated include Gradient Descent Backpropagation (traingd), Gradient Descent with Adaptive Learning Rate (traingda), and Gradient Descent with Momentum and Adaptive Learning Rate (traingdx). The dataset used is the average wholesale rice price in Indonesia, obtained from the Central Statistics Agency (BPS) website. The evaluation results show that the traingdx learning function with a 5-5-1 architecture model achieves the highest accuracy of 83.33%, representing an 8.3% improvement over the traingd and traingda learning functions, which both achieved a maximum accuracy of 75%. Based on this study, it can be concluded that using various learning functions in Backpropagation yields better accuracy compared to standard Backpropagation
Optimizing Multilayer Perceptron with Cost-Sensitive Learning for Addressing Class Imbalance in Credit Card Fraud Detection
No full textThe increasing use of credit cards in global financial transactions offers significant convenience for consumers and businesses. However, credit card fraud remains a major challenge due to its potential to cause substantial financial losses. Detecting credit card fraud is a top priority, but the primary challenge lies in class imbalance, where fraudulent transactions are significantly fewer than non-fraudulent ones. This imbalance often leads to machine learning algorithms overlooking fraudulent transactions, resulting in suboptimal performance. This study aims to enhance the performance of Multilayer Perceptron (MLP) in addressing class imbalance by employing cost-sensitive learning strategies. The research utilizes a credit card transaction dataset obtained from Kaggle, with additional validation using an e-commerce transaction dataset to strengthen the robustness of the findings. The dataset undergoes preprocessing with RUS and SMOTE techniques to balance the data before comparing the performance of baseline MLP models to those optimized with cost-sensitive learning. Evaluation metrics such as accuracy, recall, F1 score, and AUC indicate that the optimized MLP model significantly outperforms the baseline, achieving an AUC of 0.99 and a recall of 0.6. The model's superior performance is further validated through statistical tests, including Friedman and T-tests. These results underscore the practical implications of implementing cost-sensitive learning in MLPs, highlighting its potential to significantly enhance fraud detection accuracy and offer substantial benefits to financial institutions
Enhanced Yolov8 with OpenCV for Blind-Friendly Object Detection and Distance Estimation
Get PDFThe development of computer technology and computer vision has had a significant positive impact on the daily lives of blind people, especially in efforts to improve their navigation skills. This research aims to introduce a superior object detection method, especially to support the sustainability and effectiveness of blind navigation. The main focus of the research is the use of YOLOv8, the latest version of YOLO, as an object detection method and distance measurement technology from OpenCV. The main challenge to address involves improving object detection accuracy and performance, which is an important key to ensuring safe and effective navigation for blind people. In this context, blind people often face obstacles in their mobility, especially when walking in environments that may be full of obstacles or obstacles. Therefore, better object detection methods become essential to ensure the identification of nearby objects that may involve obstacles or potential threats, thus preventing possible accidents or difficulties in daily commuting. Involving YOLOv8 as an object detection method provides the advantage of a high level of accuracy, although with a slight increase in detection duration and GPU power consumption compared to previous versions. The research results show that YOLOv8 provides a low error rate, with an average error percentage of 3.15%, indicating very optimal results. Using a combined performance evaluation approach of YOLOv8 and OpenCV distance measurement metrics, this research not only seeks to improve accuracy but also efficiency in detection time and power consumption. This research makes an important contribution to the presentation of technological solutions that can help improve mobility and safety for blind people, bringing a real positive impact on the facilitation of their daily lives
ESP32 and MAX30100 with Chebyshev Filter for Enhanced Heart and Oxygen Measurement
No full textHealth monitoring is important in the technology and information era. A health monitoring device must possess high accuracy in monitoring an individual's health. The MAX30100 sensor still exhibits low accuracy and requires improvements to enhance its precision. This study proposes a remote health monitoring system based on a MAX30100 sensor for heart rate and oxygen saturation detection. The digital signal processing method uses the Chebyshev II filter on PPG to reduce noise, and the RSA algorithm is employed to enhance data security. The results of testing the MAX30100 sensor value without a filter produced the lowest error value of 0.97%, the highest 6.59% for BPM, the lowest error value of 1.88%, and the highest error of 2.66% for SpO2. The MAX30100 sensor with the Chebyshev II filter that the author proposed has the highest level of accuracy with a low error value compared to previous tests, with the lowest error value of 0.23% and the highest 0.99% for BPM and the lowest error value of 0% and the highest error of 0.2% for SpO2. The RSA algorithm ensures secure data transmission from data modification by eavesdroppers. The average total time required by the system is 542.9 ms
Improved Backpropagation Using Genetic Algorithm for Prediction of Anomalies and Data Unavailability
No full textAnomalies and data unavailability are significant challenges in conducting surveys, affecting the validity, reliability, and accuracy of analysis results. Various methods address these issues, including the Backpropagation Neural Network (BPNN) for data prediction. However, BPNN can get stuck in local minima, resulting in suboptimal error values. To enhance BPNN's effectiveness, this study integrates Genetic Algorithm (GA) optimization, forming the BPGA method. GA is effective in finding optimal parameter solutions and improving prediction accuracy. This research uses data from the 2022 National Socio-Economic Survey (Susenas) in Solok District to compare the prediction performance of BPNN, Multiple Imputation (MI), and BPGA methods. The comparison involves training the models with a subset of the data and testing their predictions on a separate subset. The BPGA method demonstrates superior accuracy, with the lowest mean squared error (MSE) and highest average accuracy, outperforming both BPNN and MI methods
Forecasting the Stock Price of PT Astra International Using the LSTM Method
No full textStocks are one of the long-term investment options and represent ownership in a company that can be acquired through buying and selling. Investment carries both the profit potential and the risks that investors must face when providing their capital to companies. Accurate stock price forecasts are very important because they provide an estimate of risk. This research aims to forecast the stock price of PT Astra International Tbk (ASII.JK) using a Long Short-Term Memory (LSTM) method. Data set closing stock prices were taken from January 2, 2015, to December 30, 2020, with a total observation of 1506. This data set is divided into 80% for training and 20% for training. The forecasting results show that the best performances have MSE, MSE, MAE and MAPE are 151.910, 23076.561, 118.128, and 2.3%, respectively. The model has a batch size of 4 and epochs of 50. This research recommends that other parties consider this method when they need to manage their investment risk in stocks
Improving Performance of KNN and C4.5 using Particle Swarm Optimization in Classification of Heart Diseases
No full textHeart disease is a major problem that must be overcome for human life. In recent years, the volume of medical data related to heart disease has increased rapidly, and various heart disease data has collaborated with information technology such as machine learning to detect, predict, and classify diseases. This research aims to improve the performance of machine learning classification methods, namely K-Nearest Neighbor (KNN) and Decision Tree (C4.5) with particle swarm optimization (PSO) feature in cases of heart disease. In this research, a comparison was made of the performance of the PSO-based K-NN and C4.5 algorithms. Following experiments employing PSO optimization to improve the K-NN and C4.5 algorithms, the findings indicated that the K-NN algorithm performed exceptionally well with PSO, achieving an accuracy of 89.09%, precision of 89.61%, recall of 90.79%, and an AUC value of 0.935