| 摘要: |
| 针对PM2.5传统点值预测会损失浓度值的波动信息，进而无法充分表示和估计其波动和变化的区间范围，提出了一种基于长短期记忆模型(LSTM)-混合模糊时间序列(HFTS)-误差修正(EC)的PM2.5区间多尺度组合预测方法；在结合深度学习和区间多尺度分解方法的基础上，进一步考虑预测误差中隐含的有效信息，建立区间时间序列组合预测模型；该模型能够从随机性较大的时间序列中提取复杂数据特征，解决传统预测方法存在的滞后性以及对误差信息利用不充分等问题；最后，通过实证分析说明该方法适用于具有较大波动的PM2.5区间预测，与已有方法相比具有较高的精确度和良好的适用性。 |
| 关键词: 区间组合预测 PM2.5 长短期记忆神经网络 误差修正 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Research on PM2.5 Interval Multi scale Combination Prediction Based on LSTM HFTS-EC |
|
LUO Rui, LIU Jin-pei, CHEN Hua-you, TAO Zhi-fu1,2
|
|
1.Business School, Anhui University,Hefei 230601, China;2.School of Mathematical Science,Anhui University, Hefei 230601, China
|
| Abstract: |
| Traditional PM2.5 point value prediction would lose the fluctuation information of concentration value, and thus could not adequately represent and estimate the range of its fluctuation and change. A multi scale combination prediction method for PM2.5 range is proposed based on long short term memory (LSTM), hybrid fuzzy time series (HFTS) and error correction (EC). Based on deep learning and interval multi scale decomposition method, the combined prediction model of interval time series is established by further considering the effective information hidden in the prediction error. This model can extract complex data features from time series with large randomness, and solve the problems of lag existing in traditional forecasting methods and insufficient use of error information. Finally, the empirical analysis shows that this method is suitable for the prediction of PM2.5 range with large fluctuation, and has higher accuracy and good applicability by comparing with the existing methods. |
| Key words: interval combination forecast PM2.5 long short term memory error correction |
