OBL-SSA optimization-based LSTM manual monitoring time sequence data generation method

The invention relates to the technical field of hydropower engineering, in particular to an OBL-SSA optimization-based LSTM manual monitoring time sequence data generation method. According to the method, a recurrent neural network with four interaction layers in a repetitive network is adopted. The method not only can extract information from sequence data like a standard recurrent neural network, but also can retain information with long-term correlation from previous farther steps. The OBL-SSA optimized LSTM has enough long-term memory to process a monitoring curve, the sampling interval is relatively small, and long-term correlation exists in the monitoring curve. Firstly, the OBL-SSA algorithm is used for preprocessing the time series data, then the LSTM algorithm is used for generating the time series data, and the problems that the observation data sequence distance interval is too large and the data correlation is poor are well solved. 本发明涉及水电工程技术领域，具体说是一种基于OBL-SSA优化的LSTM人工监测时序数据生成方法。本方法采用一种在重复网络中具有4个