The Research On Target Threat Estimation Based On Gradient Boosting Decision Tree And Back Propagation Neural Network

Rapid and effective evaluation of enemy targets is one of the core functions of the warship operational command system,which is directly related to the decision made by the commanders.To solve this problem,a target threat assessment model based on the threat of naval air targets,using gradient boosting decision tree(GBDT)and back propagation neural network(BP)algorithm,is established to study target threat assessment.The main structure of this paper is described as follows:Firstly,this article explores the target threat assessment in the status of information fusion,summarizes the current research of target threat assessment at home and abroad,analyzes the methods adopted by domestic and foreign scholars in target threat estimation,introduces the steps of threat assessment in detail,and establishes the target threat assessment model according to the steps.What's more,aiming at the characteristics of target threat assessment,this paper puts forward the target threat estimation evaluation index.Secondly,the gradient boosting decision tree algorithm is introduced to target threat estimation,and the gradient boosting decision tree algorithm is improved.A target threat assessment method based on improved gradient boosting decision tree algorithm is proposed.The effectiveness of the algorithm is verified by experiments,and the error of the method is obviously smaller than the random forest algorithm and MPSO-BP.Finally,directed against the shortcomings of BP neural network,like slow convergence speed and easy to fall into the local optimal solution,BP neural network is improved,and differential evolution algorithm is added to improve the efficiency and performance of BP neural network.The algorithm preserves the diversity of the data to a great extent and avoids the interference of human factors.The experimental results show that the BP neural network algorithm optimized by differential evolution algorithm can be used for target threat assessment.