Research Of Atmospheric Temperature Prediction Based On Deep Learning

Weather Forecasting is a challenge task with great significance, and it’s the common product of economic and social development needs and meteorological science and technology development. Atmospheric temperature prediction is not only the key point but also the difficulty of weather forecasting. Owing to its unusual effect on the social life and production, atmospheric temperature prediction has became one of the topics who have the most attention. With the fast development of the technology of the scientific meteorological data observation, the meteorological data is showing explosive growth, and the traditional forecasting methods of machine learning not only can’t adapt to the volume of massive data, but also can not obtain the representable feature. At the same time, the generalization ability of the also has a very obvious deficiencies facing the complex problems. The advent of large meteorological data, has brought new opportunities and challenges to the weather forecasting which is encountering bottlenecks in the development. For the above, this paper proposes a method based on the deep learing method which can be adapt to the mass data to make the meteorological temperature prediction research and exploration. The deep learning models and atmospheric temperature prediction theory is introduced in this article, and the feasibility is proved by experiment of the deep learning models use to the atmospheric temperature prediction. And in contrast to the shallow neural networks and support vector machines, the deep learning models get better performance. It provides a theoretical basis for the application and development of atmospheric temperature prediction using the deep learning methods.