一种基于ISSA-SVR融合模型的战时装备器材预测方法OA

Accurate prediction of equipment spare parts is crucial to ensure military equipment availability and sustained combat capability.However,current wartime forecasting faces challenges including scarce historical samples,highly nonlinear patterns,and significant uncertainty.To address these challenges,this study proposes an ISSA-SVR fusion model with three key steps:First,constructing a damage indicator system quantified via fuzzy comprehensive evaluation.Second,augmenting limited datasets using Conditional Tabular Generative Adversarial Networks(CTGAN).Third,enhancing the Sparrow Search Algorithm(SSA)with cross-directional crossover and elite retention strategies.Experimental results demonstrate a maximum relative error of merely 3%,with prediction accuracy and stability that significantly outperforming outperforms methods.This approach provides effective technical support for wartime equipment spare parts forecasting.