Reservoir level prediction and early warning method and system based on mixed deep learning model

The invention relates to a reservoir level prediction and early warning method and system based on a mixed deep learning model. The method comprises the following steps: 1) constructing a Conv1D-LSTM optimization hybrid model based on an attention mechanism and an improved particle swarm: obtaining reservoir rainwater condition data as model training data, constructing the Conv1D-LSTM hybrid model based on the attention mechanism, and then optimizing the Conv1D-LSTM hybrid model by adopting an improved particle swarm optimization algorithm; 2) loading an optimized hybrid model based on an attention mechanism and an improved particle swarm Conv1D-LSTM (Long Short Term Memory); real-time reservoir rainwater condition monitoring data are obtained and then input into the model for reservoir water level prediction, and a reservoir water level prediction value is obtained; and 3) judging whether the predicted water level is in a safety interval, if so, normally carrying out reservoir water level scheduling, otherwi