Search

Due to the outstanding strength of advanced machine-learning techniques, they have become increasingly common in predictive studies in recent years, particularly in predicting ship energy performance. In constructing predictive models, prior studies have mostly employed vessels' technical parameters to establish machine-learning algorithms. To bridge this research gap and enable wider applications, this paper presents the design of a multilayer perceptron artificial neural network (MLP ANN) as a machine-learning technique to estimate ship fuel consumption. We utilized the real operational data from 100–143 container ships to estimate fuel consumption for five different container ships grouped by size. We compared the performance of two ANN models and two multiple-regression models. Four input parameters (sailing time, speed, cargo weight, and capacity) were included in the first ANN and the first regression model, while the other two models only consider two inputs from physical function. The mean absolute percentage error of the ANN models with four inputs was the smallest and less than those in extended statistical models, demonstrating the MLP's superiority over the statistical model. The MLP ANN model can thus be applied to confirm the effectiveness of the slow-steaming method for achieving energy efficiency. [ABSTRACT FROM AUTHOR]

In South Korea, rice is the most important grain crop that it is crucial to develop yield estimation model for supporting sustainable agriculture and national food security. The main objectives of this paper are (1) to propose a Deep Learning (DL) algorithm, the Stacked Sparse Auto-encoder (SSAE), for rice-yield estimation using climatic and Moderate Resolution Imaging Spectroradiometer (MODIS) data; (2) to choose scenarios showing the best combined performance in terms of length of crop season (before and after harvest) and aggregation periods (7, 10, 15, and 30 days) of climatic data; and (3) to analyse the results in both the temporal and spatial perspectives. In this procedure, the SSAE model was built around the study objectives and compared with the artificial neural network (ANN) model for evaluation of its performance. According to the results, the combined 15-day-aggregated climatic data (between June and August) and MODIS data were selected as the optimal feature set for rice-yield estimation in the study area, Jeolla-do; the SSAE model outperformed ANN, showing root mean square error (RMSE) and RMSE% of 33.09 kg(10a)−1 (5.21 kg(10a)−1lower than ANN) and 6.89% (1.14% lower than ANN). Throughout the experiments, the rice-yield-estimation potentiality of a DL algorithm, namely the SSAE, was substantiated. [ABSTRACT FROM AUTHOR]