Ridge regression ensemble of machine learning models applied to solar and wind forecasting in Brazil and Spain

Carneiro, Tatiane C.; Rocha, Paulo Alexandre Costa; Carvalho, Paulo César Marques de; Fernández‐Ramírez, Luis M.

doi:10.1016/j.apenergy.2022.118936

Cited by 76 publications

(26 citation statements)

References 64 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MLP uses activation functions to find mathematical relationships between inputs and outputs. In the data set of nonlinear relationship, nonlinear activation functions such as logistic function (Sigmoid) and hyperbolic tangent function (tanh) are often used …”

Section: Proposed Methodologymentioning

confidence: 99%

See 1 more Smart Citation

Yield and Properties Prediction Based on the Multicondition LSTM Model for the Solvent Deasphalting Process

et al. 2023

View full text Add to dashboard Cite

Solvent deasphalting (SDA) is a complex multiscale continuous process. The operation mode of the SDA process is not considered in the related data-driven model. Therefore, this paper proposes a time lag process prediction model with multiple operation modes to solve the above problem. First, based on random forests, the relative importance of initial input variables in the SDA process on DAO yield and Conradson carbon residual are studied and features are selected according to the results. Then, the stack denoising autoencoder (SDAE) is used to reconstruct the data and obtain the nonlinear mapping information of hidden layers of SDAE and achieve feature dimension reduction. SDAE can improve clustering accuracy of fuzzy c-means, and the operation mode of SDA process is accurately divided. Long short-term memory (LSTM) is used to establish a multicondition LSTM model. Compared with the traditional LSTM model, the multicondition LSTM model has a higher prediction accuracy with R 2 > 0.95. The sensitivity analyses of the properties of feed and operating conditions on DAO yield are consistent with the principle of two-phase countercurrent extraction in the SDA process. In addition, the benchmark test of the Tennessee Eastman process shows that the proposed method is also effective in the fault detection of other processes. Because the multicondition LSTM can predict the future process measurement data according to operating mode, it can better avoid the false alarm problem and predict the fault earlier.

show abstract

Section: Proposed Methodologymentioning

confidence: 99%

“…In the data set of nonlinear relationship, nonlinear activation functions such as logistic function (Sigmoid) and hyperbolic tangent function (tanh) are often used. 43 2.6. Model Training Strategy and Testing.…”

Section: Fuzzy C-meansmentioning

confidence: 99%

Yield and Properties Prediction Based on the Multicondition LSTM Model for the Solvent Deasphalting Process

et al. 2023

View full text Add to dashboard Cite

show abstract

“…Given a univariate time series , choose a value as the time lag. This time series is transformed into a new dataset that can be used for supervised learning, which form can be expressed as follows: (12) Divide the new dataset according to the ratio of about 8:1:1, and the divided datasets are the training, validation, and test sets, respectively. First, use the WOA algorithm to find the optimal hyperparameters of the model on the training set, then use the validation set to validate the hyperparameters.…”

Section: Complete Multi-step Forecasting Scheme Based On Woa-lassomentioning

confidence: 99%

“…In the face of the high-dimensional massive dataset of this paper on natural gas, selecting the method for feature dimensionality reduction [10] is also particularly important. Traditional feature selection methods such as stepwise regression [11], ridge regression methods [12], and principal component regression can only achieve some of these objectives. In order to find a feasible solution to this problem, Tibshirani proposed in 1996 a method called Bridge Regression, inspired by Frank's Bridge Regression [13] and Bireman's Nonnegative Garrote Tibshirani proposed a new variable selection method called LASSO [14].…”

Section: Introductionmentioning

confidence: 99%

Forecasting Study of Natural Gas Consumption by Combined Models Based on LASSO and WOA

Tang

Qian

2023

ARJOM

View full text Add to dashboard Cite

As the impact of the Russia-Ukraine war continues to expand, energy shortages appear in Europe. After Russia cut off the Nord Stream 1 pipeline that transported natural gas to Europe, most European countries experienced a natural gas crisis, severely affecting Germany. In order to effectively predict the consumption of natural gas, this paper combines the Least Absolute Shrinkage and Select Operator model with the Whale Optimization Algorithm, uses the NAR model to reconstruct the phase space of the original time series, and performs a 5-step forward forecast. Use the model to forecast a German monthly natural gas consumption dataset. Comparing the results of WOA-LASSO with other five other WOA-based hybrid models and Cross-Validation based models for prediction results, it is found that WOA-LASSO has the smallest MAPE in each step of the 5-step prediction, and the numerical results are between 8.273% and 9.867%. Moreover, when comparing WOA with the conventional optimization scheme Cross-Validation, it is found that WOA can obtain better model hyperparameters, which can effectively enhance the generalization performance and prediction accuracy of the model.

show abstract

“…Ensemble learning has been shown to be effective in as varied uses as forecasting solar and wind availability (Carneiro et al 2022) and optimizing genetic transfer in chrysanthemum using Agrobacterium (Hesami et al 2020) to, in some cases, genomic prediction of phenotypes (Azodi et al 2019;Liang et al 2021). In addition to the examples above, ensemble methods have been used within agriculture to predict soybean yield based on measured yield component traits (Yoosefzadeh-Najafabadi et al 2021b).…”

Section: Introductionmentioning

confidence: 99%

Ensemble of BLUP, Machine Learning, and Deep Learning Models Predict Maize Yield Better Than Each Model Alone

Kick

Washburn

2023

Preprint

View full text Add to dashboard Cite

Predicting phenotypes accurately from genomic, environment, and management factors is key to accelerating the development of novel cultivars with desirable traits. Inclusion of management and environmental factors enables in silico studies to predict the effect of specific management interventions or future climates. Despite the value such models would confer, much work remains to improve the accuracy of phenotypic predictions. Rather than advocate for a single specific modeling strategy, here we demonstrate within large multi-environment and multi-genotype maize trials that combining predictions from disparate models using simple ensemble approaches most often results in better accuracy than using any one of the models on their own. We investigated various ensemble combinations of different model types, model numbers, and model weighting schemes to determine the accuracy of each. We find that ensembling generally improves performance even when combining only two models. The number and type of models included alter accuracy with improvements diminishing as the number of models included increases. Using a genetic algorithm to optimize ensemble composition reveals that, when weighted by the inverse of each model’s expected error, using combinations of best linear unbiased predictors, linear fixed effects models, deep learning models, and select machine learning models perform best on our datasets.

show abstract

Ridge regression ensemble of machine learning models applied to solar and wind forecasting in Brazil and Spain

Cited by 76 publications

References 64 publications

Yield and Properties Prediction Based on the Multicondition LSTM Model for the Solvent Deasphalting Process

Yield and Properties Prediction Based on the Multicondition LSTM Model for the Solvent Deasphalting Process

Forecasting Study of Natural Gas Consumption by Combined Models Based on LASSO and WOA

Ensemble of BLUP, Machine Learning, and Deep Learning Models Predict Maize Yield Better Than Each Model Alone

Contact Info

Product

Resources

About