Prediction of World Temperature Based on PSO Optimized LSTM Neural Network

Global temperature prediction is the crucial to prevent global warming. For the global ballooning problem, this paper proposes a long short-term memory (LSTM) neural network model combined with particle swarm optimization algorithm for predicting global average temperature. Two metrics, mean square error (MSE), mean absolute percentage error (MAPE), are selected to evaluate the prediction performance. The prediction results are compared with BP, LSTM, and PSO-BP. Particle swarm optimization is used to optimize the LSTM network input weights, reduce the loss function, and fit faster. Sensitivity analysis of the model yields a good generalization of the model, which can be extended for time series prediction.