Research On Lightning Disaster Text Clustering And Prediction Method Based On Enhanced Gray Wolf Optimization Algorithm

Environmental meteorology is closely related to our production and life.While meteorology brings us convenience,it also brings some negative effects,such as thunderstorm disaster.As one of the meteorological disasters that have the most extensive impact on our life,thunderstorm always threatens the safety of people's life and property.Therefore,how to rationally utilize the historical thunderstorm disaster data to discover the potential information contained in the meteorological data;It has become the focus of scientific research to predict lightning disaster accurately according to appropriate meteorological factors.At present,the clustering method of lightning disaster text still has the problem that the clustering results are not reliable,which leads to the failure to dig out deeper information.Moreover,the method of lightning disaster prediction also has the problems of low accuracy and high time complexity of prediction algorithm.Therefore,in view of the above problems,this thesis carried out related research on lightning disaster text clustering and prediction based on the enhanced gray Wolf optimization algorithm.The main research contents of the subject are as follows:(1)Aiming at the problem of lightning disaster text clustering,a k-means lightning disaster text clustering algorithm based on the enhanced gray wolf optimization algorithm was proposed.The algorithm classified the lightning disaster types and put forward effective processing strategies,which effectively overcame the shortcomings of inaccurate clustering results of traditional k-means algorithm.Firstly,the text data of lightning disaster is converted into numerical data through text vectorization operation.Secondly,the immune cloning operation was used to select the elite individuals from the gray wolf population to form the elite population,and the mutation operation was performed on the population to enrich the characteristics of the elite gray wolf individuals and reduce the possibility of premature convergence of the enhanced gray wolf optimization algorithm.Secondly,the idea of particle swarm position updating was used to consider the position information of a single elite gray wolf.Finally,enhanced gray wolf optimization algorithm is used to find the optimal clustering center of K-means algorithm for text clustering analysis.Simulation results show that the proposed algorithm has better clustering accuracy,recall rate and F value than the existing algorithms,and the text clustering effect is more referential.Therefore,the study of thunderstorm disaster text clustering using this algorithm can better find the information contained in the text of lightning disaster.(2)Aiming at the problem of lightning disaster prediction,a lightning disaster prediction model based on enhanced gray wolf optimization algorithm to optimize BP neural network was proposed,which solved the problem of inaccurate prediction results of traditional lightning disaster prediction methods.Firstly,the weights and thresholds of BP neural network were initialized,and the weights and thresholds were set as each individual in the gray wolf population.Secondly,the fitness function of the enhanced gray wolf optimization algorithm is defined as the mean square error of BP neural network,and the enhanced gray wolf optimization algorithm is used to search for the individual gray wolf to get the most suitable weight and threshold value.Finally,the weights and thresholds were used to construct a network model for network training,and the classical UCI data set was used to evaluate the algorithm performance of the enhanced gray wolf optimization algorithm for optimizing BP neural network model.The simulation results show that,compared with the original BP algorithm and the common improved BP algorithm,the proposed algorithm can converge more quickly,and the algorithm has higher accuracy and better stability.Finally,based on the real data of lightning disaster,the prediction results show that the proposed algorithm has higher prediction accuracy and lower null rate of lightning disaster.