Search

Due to climatic change, a variation in meteorological aspects influences the water requirement for crops, evapotranspiration, and water allocation of agro-meteorological and agriculture. Accurate estimation of Evapotranspiration ( ET o ) has great importance to improve the utilization of water efficiently and irrigation scheduling. The main overarching goal of this paper is to investigate the abilities and applicability of three supervised machine learning models: Extreme Machine Learning ( ELM 1 , ELM 2 , ELM 3 , ELM 4 ), Multi-layer Perceptrons-Neural Network ( MLP 1 , MLP 2 , MLP 3 , MLP 4 ), Support Vector Machine ( SVM 1 , SVM 2 , SVM 3 , and, SVM 4 ) to modeling the daily ET o . Further, a three-layer multi-model ensemble machine learning approach is presented to predict evapotranspiration ET o . The first layer consists of different statistical models to produce individual forecasts. The blending approach is employed to create an ensemble of the forecasts generated by the initial layer to produce probabilistic forecasts. In the second layer, three ensemble models ( Ensemble ELM , Ensemble MLP , Ensemble SVM ) are trained for prediction of ET o by using the previous layer predictions and training data. In the third-layer, accuracy of the ( ET o ) is estimated by tuning the parameters of second layer ensemble model. It has been analyzed that all statistical models showed effectiveness in high performance for modeling everyday ET o (e.g. Nash-Sutchliffe efficiency (NSE)= 0.93-0.99, coefficient of determination (r 2 ) = 0.93-0.99, Accuracy (ACC) = 80-99, Mean Square error (MSE) = 0.0103-0.1516). Particularity, the ensemble method with SVM achieved good accuracy (99.46% to 99.72%) to predict the daily ET o and correlation coefficient is closed to 1 on training, validation, and testing datasets. Its root means square error (RMSE) (0.0085 to 0.0935) and Mean Absolute Error (MAE) (0.0614 to 0.0639) are minimum as compared to other ensemble machine learning models. [ABSTRACT FROM AUTHOR]

The precision of reference evapotranspiration (ETo) predictions would vary, depending on the adopted empirical method and the availability of meteorological data. This study aims to enhance the prediction accuracy of ETo using the high-order response surface method (HO-RSM). Daily scale climatological information are used to build the predictive model including maximum temperature (Tmax), maximum humidity (Hmax), wind speed (WS), solar radiation (SR), and vapor pressure deficit (VPD), which are obtained from three observation stations in Burkina Faso, West Africa. Ten models corresponding to ten different input combination sets are evaluated for variability influence by comparing the predicted ETo with the observed ETo. The models presented a similar performance at both Gaoua and Boromo stations with the determination coefficient (R2) and root mean square error (RMSE) values ranging between 0.6831–0.9966 (0.0622–0.5065) and 0.7237–0.9948 (0.0722–0.4942), respectively. As for the Dori station, the models showed a lower performance with R2 (RMSE) values ranging between 0.2068 and 0.5229 (0.8292–1.0051), which may be due to the insufficient input variables or the requirement of higher order in RSM modeling for this station. Results also showed that the M10 model that includes all five input variables performed the best at three stations, with respect to the statistical performance. This is followed by the M7 model, which excluded the Hmax in the prediction, suggesting that Hmax has the least influence on the ETo prediction among all the input variables. The insignificant trend in selecting the optimum order of the RSM also showed that HO-RSM is case sensitive and hence precautions are required for generalizing model applications. [ABSTRACT FROM AUTHOR]

Highlights • Estimation of the reference evapotranspiration (ETo) of Punjab. • Different H 2 O-based machine learning models are compared in predicting daily ETo. • The performance of DL model is compared to Penman–Monteith, RF, GLM and GBM models. • DL model performed better than the considered models for training, validation and testing sets. Abstract Over the last decade, the combination of both big data and machine learning research area's receiving considerable attention and expedite the prospect of the agricultural industry. This research aims to gain insights into a state-of-the-art big data application in smart farming. An essential issue for agriculture planning is to estimate evapotranspiration accurately because it plays a pivotal role in irrigation water scheduling for using water efficiently. This article presents H 2 O model framework to determine the daily ET o for Hoshiarpur and Patiala districts of Punjab. The effects of four supervised learning algorithms: Deep Learning-Multilayer Perceptrons (DL), Generalized Linear Model (GLM), Random Forest (RF), and Gradient-Boosting Machine (GBM) and also evaluate the overall ability to predict future ET o . Analysis of these four models, perform in H 2 O framework. This framework presents a new criterion to train, validate, test and improve the classification efficiency using machine learning algorithms. The performance of the DL model is compared with other state-of-art of models such as RF, GLM and GBM. In this respect, our analysis depicts that models presents high performance for modeling daily ET o (e.g. NSE = 0.95–0.98, r2 = 0.95–0.99, ACC = 85–95, MSE = 0.0369–0.1215, RMSE = 0.1921–0.2691). [ABSTRACT FROM AUTHOR]

• We proposed Fuzzy-Genetic and Regularization Random Forest models for modeling the agriculture application. • A novel multi-layer ensemble model is proposed for predicting the Kc and ETc of Ludhiana station. • Our model results show the developed method have high precision. Smart farming has played a significant role in decision support system to maximize the yield with minimum consumption of water in the field of agriculture. The main objective of this paper is to design and develop an innovative multilevel model ensembling for accurate estimation of crop coefficient (K c) and reference evapotranspiration (ET c) using Fuzzy-Genetic (FG) and Regularization Random Forest(RRF) models. This study present the water requirement of three crops namely (maize, wheat 1 and wheat 2) in which ET c is a function of the product of the crop coefficient K c and reference evapotranspiration (ET o). The proposed model is used to analyze the data collected by IMD, Pune and PAU, Ludhiana (case study) for decision making in a crop water model. The proposed FG-RRF(ET c) crop prediction model efficiently estimated K c and ET c and make an efficient decision. [ABSTRACT FROM AUTHOR]