This paper presents a CNN-LSTM deep neural network for meteorological data. At first, the theory of model is explained. Then, The thunderstorm prediction model of Beijing in the next 6 hours was established with the observation data(every 6 hours) of in Beijing(station no. 54511) and ECMWF reanalysis data (http://apps.ecmwf.int/datasets/data/interim-fulldaily/levtype=pl/) on January 1, 1984 to September 14, 2011. We tested the model for the forecast accuracy of the thunderstorm of the city of Beijing, on September 15, 2013 to December 30, 2013.The results show that the model has an accuracy of 83.08%, the rate of false positives 13.02% for the prediction of thunderstorm in Beijing in the next 6 hours, can provide good support for the 6-hour proximity forecast of thunderstorm.
. A WEATHER PREDICTION MODEL BASED ON CNN AND RNN DEEP NEURAL NETWORK——AN EXAMPLE IS THE 6 HOUR FORECAST OF THUNDERSTORM IN BEIJING[J]. Journal on Numerical Methods and Computer Applicat, 2018, 39(4): 299-309.
Delden A V. The synoptic setting of thunderstorms in western Europe[J]. Atmospheric Research, 2001, 56(1):89-110.
Eck D, Schmidhuber J. Learning the Long-Term Structure of the Blues[J]. Lecture Notes in Computer Science, 2002, 2415:284-289.
Dahl G E, Yu D, Deng L, et al. Context-Dependent Pre-Trained Deep Neural Networks for LargeVocabulary Speech Recognition[J]. IEEE Transactions on Audio Speech & Language Processing, 2011, 20(1):30-42.
Hinton G, Deng L, Yu D, et al. Deep Neural Networks for Acoustic Modeling in Speech Recognition:The Shared Views of Four Research Groups[J]. IEEE Signal Processing Magazine, 2012, 29(6):82-97.
Graves A. Supervised Sequence Labelling with Recurrent Neural Networks[J]. Studies in Computational Intelligence, 2008, 385.
Graves A. Long Short-Term Memory[M]//Supervised Sequence Labelling with Recurrent Neural Networks. Springer Berlin Heidelberg, 2012, 1735-1780.
Haklander A J, Delden A V. Thunderstorm predictors and their forecast skill for the Netherlands[J]. Atmospheric Research, 2003, 67-68(03):273-299.
Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems. Curran Associates Inc. 2012, 1097-1105.
Kingma D, Ba J. Adam:A Method for Stochastic Optimization[J]. Computer Science, 2014.
Misra S, Sarkar S, Mitra P. Statistical downscaling of precipitation using long short-term memory recurrent neural networks[J]. Theoretical & Applied Climatology, 2017, (5):1-18.
Srivastava N, Hinton G, Krizhevsky A, et al. Dropout:a simple way to prevent neural networks from overfitting[J]. Journal of Machine Learning Research, 2014, 15(1):1929-1958.
Pang L, Lan Y, Guo J, et al. Text Matching as Image Recognition[J]. 2016.
Sutskever I, Vinyals O, Le Q V. Sequence to Sequence Learning with Neural Networks[J]. 2014, 4:3104-3112.
Shi X, Chen Z, Wang H, et al. Convolutional LSTM Network:a machine learning approach for precipitation nowcasting[C]//International Conference on Neural Information Processing Systems. MIT Press, 2015, 802-810.
Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition[J]. Computer Science, 2014.
Mccann D W. A Neural Network Short-Term Forecast of Significant Thunderstorms[J]. Weather & Forecasting, 1992, 7(3):525-534.