论文部分内容阅读
针对氧化铝配料过程波动较大的非平稳时间序列参数预测问题,提出了一种基于小波分析的ARMA(自回归移动平均)模型、BP神经网络模型和Holt-Winters非季节模型的组合模型。首先通过小波分解,将原始时间序列依尺度分解为同长度不同频率的数据,然后对经过处理的高频数据和低频数据分别采用不同的模型进行预测,最后将不同频率的所有预测数据进行重构得到原始时间序列的组合预测模型。预测结果表明,所建立的模型对波动较大的非平稳时间序列的预测具有很大的优势。
Aiming at the problem of parameter estimation of non-stationary time series with large fluctuation of alumina batch process, a combined model of ARMA (autoregressive moving average) model, BP neural network model and Holt-Winters non-seasonal model based on wavelet analysis is proposed. Firstly, the wavelet decomposition is used to decompose the original time series into different lengths of data with the same length. Then the processed high frequency data and low frequency data are respectively predicted by different models. Finally, all the prediction data of different frequencies are reconstructed Get the original time series combined forecasting model. The prediction results show that the proposed model has great advantages for the prediction of non-stationary time series with large fluctuation.