Mixed time series data prediction method based on EEMD-CEEMDAN combined with LSTM

• Main Authors: TANG HAO, GAO JINXIONG, YOU JIA, GUO DONGSHENG, LIU HUIZHOU
• Format: Patent
• Language: Chinese; English
• Published: 03.10.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention provides a mixed time series data prediction method based on EEMD-CEEMDAN combined with LSTM, and the method comprises the following steps: obtaining original time series data, and carrying out the preprocessing of the original time series data; performing signal decomposition on the preprocessed time sequence data to obtain an IMF modal component; and inputting the obtained IMF modal component into the LSTM prediction model to obtain a final prediction result. According to the prediction model provided by the invention, the prediction accuracy of the time sequence data is further improved and enhanced by combining the EEMD-CEEMDAN with the LSTM model; a multi-step signal decomposition mode is provided, target sequence data and other component sequence data in a data set are decomposed by using EEMD and CEEMDAN respectively, and the signal decomposition efficiency is improved. 本发明提供一种基于EEMD-CEEMDAN结合LSTM的混合时间序列数据预测方法，包括下述步骤：获取原始时间序列数据，并对所述原始时间序列数据进行预处理；对经预处理之后的时间序列数据进行信号分解，得到IMF模态分量；将得到的IMF模态分量输