| 以下仅列出文章摘要、提纲简介，如需获取全文阅读权限，或原创定制、长期合作，请随时联系。 微信QQ:312050216 |
 |
 |
| Decomposition-Clustering-Ensemble(DCE)learning approach In this paper, a decomposition clustering ensemble (DCE) learning approach for ACI data prediction is proposed. The method can be decomposed by EEMD, K-means clustering analysis and LSSVR. The empirical results show that the K-LSVR learning method can significantly improve the prediction performance in the prediction accuracy and robustness analysis, which is superior to other popular forecasting methods. This shows that the EEMD-LSSVR-K-LSSVR learning method is a promising method for the prediction of ACI data. The DCE model has a strong extensibility or reference value, and can be applied to many other forecasting fields. It is believed that the performance and robustness of the model can have good prediction accuracy in other forecasting fields. In addition, this paper focuses on single variable time series analysis, while other factors are not taken into account. If these factors can be introduced into the proposed DCE learning method, the performance may be better. Only 3 steps are considered, and there are no many other steps comparisons. Such a model may not perform well in other multi-steps forecasts. Data processing methods are not perfect enough so that the data may be distorted which may influent the accuracy of the model. The kernel function of LSVR is chosen only through experience. For a particular problem, the chosen kernel function may not be optimal. |
