Heart disease is one of the most common diseases in the world. The objective of this study is to aid the diagnosis of heart disease using a hybrid classification system based on the ReliefF and Rough Set (RFRS) method. The proposed system contains two subsystems: the RFRS feature selection system and a classification system with an ensemble classifier. The first system includes three stages: (i) data discretization, (ii) feature extraction using the ReliefF algorithm, and (iii) feature reduction using the heuristic Rough Set reduction algorithm that we developed. In the second system, an ensemble classifier is proposed based on the C4.5 classifier. The Statlog (Heart) dataset, obtained from the UCI database, was used for experiments. A maximum classification accuracy of 92.59% was achieved according to a jackknife cross-validation scheme. The results demonstrate that the performance of the proposed system is superior to the performances of previously reported classification techniques.

The long short‐term memory (LSTM) recurrent neural network algorithm in deep learning has demonstrated significant superiority in predicting the realized volatility (RV) of crude oil prices. However, there is no robust and consistent conclusion regarding the handling of microstructural noise from high‐frequency data during the prediction process. Therefore, this study utilizes six commonly used data decomposition methods, as documented in the literature, to address the issue of noise handling and decompose the RV series of Chinese crude oil futures. Subsequently, LSTM is integrated with the decomposed data to model and forecast RV. The empirical findings provide compelling evidence that the LSTM model based on neural networks outperforms traditional econometric models in out‐of‐sample forecasting. Furthermore, the LSTM model with data noise decomposition consistently exhibits superior out‐of‐sample prediction performance compared to the model without noise decomposition. Among the various data noise decomposition models examined, this study highlights the significant out‐of‐sample predictive power of variational mode decomposition (VMD), a nonrecursive signal decomposition method, that outperforms other methods. In the scenario of predicting one step ahead, the VMD‐LSTM model demonstrates MAE, MSE, and HMAE values of 7.5 × 10 − 2, 1.10 × 10 − 4, and 0.423, respectively.