Air strike target recognition based on a double layer random forest

DOI:	10.11809/bqzbgcxb2023.05.030
Keywords:	air defense operations; target recognition; random forest; Gini index; feature evaluation; data dimensionality reduction
Abstract:	Air strike target recognition is a key link in air defense combat command decision making. Aiming at the problem that the complex features of air strike targets can easily cause model over fitting and a low recognition accuracy, in order to improve the recognition ability of air strike targets, this paper proposes an air strike target recognition algorithm based on a double layer random forest. The algorithm evaluates and optimizes the importance of features in the first layer of the random forest by calculating the Gini index change, and then performs data dimensionality reduction and target recognition in the second layer. Compared with traditional random forests, it can improve the accuracy and stability of target recognition. The algorithm is compared and analyzed with the traditional random forest, support vector machine and PNN neural network. The simulation results show that the algorithm ensures the recognition accuracy and has a high recognition speed and recognition stability at the same time.
Issue:	Vol. 44 No. 5 (2023)
Published:	2023-05-28
PDF HTML