Crude oil prices (COP) profoundly influence global economic stability, with fluctuations reverberating across various sectors. Accurate forecasting of COP is indispensable for governments, policymakers, and stakeholders to make well-informed decisions and effectively mitigate risks. The decomposition-based hybrid models have been showing promising COP forecasting accuracy than other time series forecasting methods. Despite this fact, no systematic study has been conducted to evaluate the true potential of different decomposition-based hybrid methods employing different forecasting models to forecast the COP. Therefore, a hybrid modeling framework is developed by combining efficient decomposition techniques, namely empirical mode decomposition (EMD), ensemble EMD (EEMD), complete EEMD with adaptive noise (CEEMDAN), and variational mode decomposition (VMD) with seven statistical models, fourteen machine learning (ML) models, and six deep learning (DL) models. Further, a systematic study is conducted on the resulting decomposition-based hybrid models to find the best hybrid model for COP forecasting. Three distinct train-test data splits are employed to ensure a reliable evaluation of the models using four performance metrics. Extensive statistical analysis is conducted to identify the optimal combination of the decomposition technique and forecasting model for precise COP prediction. The results demonstrate that the proposed decomposition-based hybrid model employing VMD and Huber Regression is statistically the best method among all alternatives to forecast monthly COP. The proposed hybrid method VMD-Huber Regression improves the root mean square error (RMSE) by 21% than CEEMDAN-ARIMA, 58.31% than EEMD-Theta, 13.18% than EMD-Random Walk, and 49.44% than VMD-TBATS hybrid methods in 60–40 Train-Test split ratio.

Accurate forecasting of Crude oil price (COP) is paramount for financial markets, energy sector stakeholders, and a nation’s economy. The COP is affected by periodic, aperiodic, and sporadic factors like geopolitical events of importing and exporting countries. Therefore, accurate forecasting of COPs is imperative and challenging. Motivated by this, in this project, extensive studies are made to make point and interval forecasting of COPs employing machine learning models. Optimized deep learning models are used, and deep features are extracted. The deep features are transformed using principal component analysis (PCA). The transformed features are modelled using shallow machine learning models. Once the point forecasts are obtained, error modelling using different distributions is performed to compute the interval forecasts at different significance levels. Several alternatives are considered in optimized deep learning models, machine learning models and distribution functions to make the methodology reliable and robust. Since the machine learning models are stochastic, the simulations are repeated, and the mean values are considered from different point and interval forecasting methods for drawing reliable conclusions. Simulation results suggests that the GRU-Linear Regression model with optimized hyper-parameters provides the best in MAE and MASE. At the same time, the CNN-Huber Regressor provides the best RMSE and SMAPE than all other alternatives. .

In this paper, individual and hybrid methods are proposed employing optimized statistical and deep learning (DL) models for deterministic (point) and probabilistic (interval) forecasting of crude oil price time series. The statistical models are optimized using the Forecast package of R. To enhance the performance of DL models, a novel pruning DE-DL method is proposed, which employs the differential evolution (DE) algorithm to optimize architecture and continuous and discrete-valued hyper-parameters. The proposed DE-DL method is so generic that it can be applied to optimize different DL models for any supervised learning problem. Five DL models (LSTM, BiLSTM, GRU, CNN, and ConvLSTM) are optimized for forecasting monthly crude oil prices and hybridized with an optimized ARIMA model for developing optimized additive and multiplicative hybrid forecasting models. The effectiveness of the proposed methods is evaluated through deterministic and probabilistic forecasting measures, comparing the results with six optimized statistical models, thirteen machine learning models, five optimized DL models, and ten optimized hybrid models. It is observed from the simulation results that the proposed optimized Additive-ARIMA-GRU hybrid model provides statistically superior forecasts, and the t Location Scale distribution is more suitable than the Gaussian distribution for computing reliable prediction intervals with different significance levels.

The crude oil price (COP) has substantial implications on world economy, as it impacts industries ranging from transportation to manufacturing. Given the volatile nature of COP, accurate forecasting is very much crucial for businesses and policymakers alike. Forecasting crude oil prices is a challenging task for the complex and volatile nature of the global oil market. As a result, estimating the price of crude oil has been a challenging and crucial component of forecasting research. In this study, we employ fourteen machine learning (ML) models for predicting the weekly and daily crude oil price. To evaluate the effectiveness of ML models, four performance measure metrics are utilized, including “mean absolute scaled error (MASE), symmetric mean absolute percentage error (SMAPE), root mean square error (RMSE), and mean absolute error (MAE)”. Detailed statistical analyses of data obtained using the Wilcoxon Signed-Rank test demonstrate that the linear support vector regression (SVR) model for weekly COP data, and linear regression for daily COP data are statistically more effective in predicting COPs than other models considered. The linear regression model acquires the statistically best rank across three accuracy metrics (SMAPE, MAE, MASE) and Gradient Boosting acquires the best rank based on RMSE accuracy metrics considering both weekly and daily COP data according to the Friedman and Nemenyi hypothesis test.

Accurate prediction of crop prices assists farmers to decide the best time to sell their produce so as to get maximum benefit and assists Government for post-harvest storage and management of the produce so as to stabilize the price volatility throughout the year. At the same time, pricing of crop depends on various factors including the amount of cultivation, demand of consumers, climate, etc. Hence, the prediction of crop prices is a challenging and important problem. Inspired from this, in this study, we have proposed two additive hybrid methods (Additive-ETS-SVM, Additive-ETS-LSTM) and five multiplicative hybrid methods (Multiplicative-ETS-ANN, Multiplicative-ETS-SVM, Multiplicative-ETS-LSTM, Multiplicative-ARIMA-SVM, Multiplicative-ARIMA-LSTM) to predict the monthly retail and wholesale price of three most commonly used vegetable crops of India, namely, tomato, onion, and potato (TOP). The obtained results are compared with two most promising statistical models, three leading machine learning models and five hybrid methods existing in the literature. Extensive statistical analyses of simulation results considering mean absolute error (MAE), symmetric mean absolute percentage error (SMAPE), and root mean square error (RMSE) confirm the superiority of the hybrid methods in predicting the TOP prices.