Intelligent Recommendation Algorithm for Resource Damage Based on Deep Learning
The battlefield situation often changes,and resource damage and faults occur frequently,requiring replacement suggestions.However,manual screening is inefficient,making it difficult to develop rules based on rule screening.The rules are complex,and sample perturbations have a significant impact on the rules.In response to these needs and shortcomings,this paper constructs a set of resource replacement recommendation algorithms based on deep learning algorithms.The algorithm uses a combination of BERT and multi-layer perceptron as the encoder,and calculates similarity from three major aspects:entity name,adjacency relationship,and attribute set.Finally,a multi-layer perceptron is used to generate resource replacement scores.The experiment on the constructed battlefield resource knowledge graph shows that the constructed algorithm can achieve a TOP1 accuracy of 93%and a TOP10 accuracy of 99%,with inference time in milliseconds,meeting the requirements for accuracy and computation time of resource recommendation algorithms in the context of changing battlefield situations.
万方数据
维普
