Temperature Prediction Based On Deep Learning

With the development of science and technology,the source of meteorological data is more and more,and the amount of data accumulated is becoming larger and larger.There are abundant meteorological laws behind these massive meteorological data.If the traditional algorithm is still used to predict the weather,it can not make full use of the data,which is a great waste of data resources.Thanks to the development of deep learning,there are many network structures which can excavate the information contained in the big data.If we can apply these algorithms to the meteorological forecast,it will be very helpful to the meteorological field.Temperature prediction is the most basic and important part of weather prediction,so this paper discusses the application of deep learning in weather prediction with temperature prediction as the starting point.This paper mainly uses the random forest,the long short-term memory network,the seqtoseq₁ model based on the long short-term memory network and the seqtoseq₂ model based on the long short-term memory network to forecast the fine temperature in the next 24 hours.The prediction results of the four models are compared and analyzed.Compared with the traditional random forest algorithm,the deep learning algorithm model can show good predictive ability in the face of large meteorological data.Especially,the long short-term network structure with its strong ability to deal with time series data,can be used as the common structure of temperature refinement prediction.