Evaluation Model Of Earthquake Direct Economic Loss Based On Machine Learning Algorithm

As a destructive natural disaster,earthquake poses a great threat to people's life and property.China is located between the ring-Pacific seismic belt and the Eurasian seismic belt.For China,a densely populated country with a rapidly growing economy,the threat of earthquakes is growing.Therefore,it is of great significance and value to evaluate the direct economic loss of earthquake quickly.After a strong earthquake,the rapid assessment of direct economic losses after the earthquake is a reference for the national government at all levels to respond quickly and respond to the disaster relief arrangements,financial input and guide the social from all walks of life to play a guiding role.Earthquake direct economic loss is different from earthquake magnitude,distribution of disaster areas and other parameters,which often need to be evaluated by some methods,such as classification list method,rapid evaluation method based on earthquake damage matrix and subsequent evaluation method based on genetic neural network.With the improvement of seismic data,it is possible to mine information from seismic data.Based on historical seismic data and using machine learning algorithm,this paper mainly completes the following work:1.The background and research significance of this paper are discussed.The development of machine learning is described and the research status of earthquake economic loss at home and abroad is summarized.2.The representative factors affecting the direct economic loss of earthquake were introduced,and the original data were collected.At the same time,the sample data were analyzed and the corresponding data processing method was adopted.3.By applying the random forest algorithm,a more optimized random forest evaluation model is obtained by adjusting the super parameters and optimizing the allocation of training set,and the direct economic loss caused by earthquake is evaluated,so as to obtain the importance ranking of the direct economic loss caused by earthquake.4.The framework of deep learning algorithm is selected to establish the neural network structure.By adjusting the network,the loss function is continuously reduced until the model converges.Meanwhile,the evaluation model of deep learning algorithm is evaluated.