An Artificial Neural Network for Predicting Groundnut Yield Using Climatic Data

AgriEngineering, Journal Year: 2023, Volume and Issue: 5(4), P. 1713 - 1736

Published: Sept. 30, 2023

Groundnut, being a widely consumed oily seed with significant health benefits and appealing sensory profiles, is extensively cultivated in tropical regions worldwide. However, the yield substantially impacted by changing climate. Therefore, predicting stressed groundnut based on climatic factors desirable. This research focuses several combinations of using artificial neural networks three training algorithms. The Levenberg–Marquardt, Bayesian Regularization, Scaled Conjugate Gradient algorithms were evaluated for their performance such as minimum temperature, maximum rainfall different Sri Lanka, considering seasonal variations yield. A three-layer network was employed, comprising hidden layer. layer consisted 10 neurons, log sigmoid functions used activation function. these configurations mean squared error Pearson correlation. Notable improvements observed when Levenberg–Marquardt algorithm applying natural logarithm transformation to values. These evident through higher correlation values (0.84), validation (1.00) testing (1.00), lower (2.2859 × 10−21) value. Due limited data, K-Fold cross-validation utilized optimization, K value 5 process. application resulted (0.3724) results revealed that performs better capturing relationships between provides valuable insights into utilization yield, highlighting effectiveness emphasizing importance carefully selecting expanding modeling equation.

Language: Английский

---

Estimating solar radiation using artificial neural networks: a case study of Fiche, Oromia, Ethiopia

Cogent Engineering, Journal Year: 2023, Volume and Issue: 10(1)

Published: June 5, 2023

The precise assessment and evaluation of global solar radiation (GSR) is crucial for designing effective energy systems. However, in developing countries like Ethiopia, the cost maintenance measuring devices are inadequate. As a result, researchers have explored alternative methods such as empirical models to estimate GSR. This article proposes using artificial neural networks (ANN) predict daily monthly averaged horizontal GSR (HGSR) around Fiche town various network types. input variables were divided into training (70%) testing (30%) sets evaluate types, with sigmoid function used activation at hidden layer linear output layer. predicted mean HGSR ranges from 3.282 kWh/m2/day 6.967 4.628 kWh/m2 6.613 respectively. values obtained compared those provided by NASA observation data found be within acceptable limits. Statistical metrics MAPE, MSE, RMSE show that CFBP, FFBP, LR, EBP better types estimating HGSR, while EBP, LR HGSR. Overall, all ANN accurately In general, findings this study indicated location had promising producing electricity uses.

Language: Английский

---

Multi-Scale Convolutional Echo State Network With an Effective Pre-Training Strategy for Solar Irradiance Forecasting

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 13442 - 13452

Published: Jan. 1, 2024

In this article, a new kind of neural network model named multi-scale convolutional echo state (MCESN) is proposed for solar irradiance prediction, which integrates the strong feature extraction capability (CNN) and fast yet efficient prediction ability (ESN). Firstly, information at different time scales (one dimensional series) data are extracted selected by CNN (MCNN) in pre-training stage. Then, trained features above concatenated passed to ESN module as input signal, can be further encoded into high-dimensional space; Meanwhile, target value fitted predicted phase. Finally, effectiveness MCESN evaluated hourly prediction. experiment, RMSE, MAE, MAPE R chosen four metrics evaluate performance model. Simulation results demonstrate that perform better than classical ESN, MCNN, backpropagation (BP) random forest (RF), long short memory (LSTM) deep (DESN) algorithms.

Language: Английский

---

Enhancing solar radiation prediction accuracy: A hybrid machine learning approach integrating response surface method and support vector regression

Ain Shams Engineering Journal, Journal Year: 2024, Volume and Issue: unknown, P. 103034 - 103034

Published: Sept. 1, 2024

Language: Английский

---

An inexpensive phytoremediation system for treating 50,000 L/day of sewage

International Journal of Phytoremediation, Journal Year: 2022, Volume and Issue: 25(8), P. 1029 - 1041

Published: Oct. 20, 2022

The paper describes the setting up and long-term continuous operation of first real-life, pilot scale, sewage treatment plant based on recently patented phytoremediation technology, trademarked as SHEFROL®. unit was about three times cheaper to install, operate maintain than least expensive other wetland-based technologies presently in vogue. Its semi-permanent version is 30 cheaper. Monitoring flow rates levels intermittently over a 3 year course indicated constancy robustness reactor treating total solids, suspended chemical oxygen demand, biological Kjeldahl nitrogen, soluble phosphorous average extents 94, 84, 79, 70, 62 28% respectively. Earlier experience with bench-scale SHEFROL® units has that removal metals like Cu, Ni, Co, Zn, Mn also takes place extent 25–45% these systems. These primary, secondary, tertiary treatments occurred single process no necessity any pumping, aeration, or recycling. Models artificial intelligence were developed which enable forecasting performance terms secondary treatment,

Language: Английский

---