Research on Multivariate Dynamic Cost Standard Prediction Model Based on Random Forest and LSTM Combined with XGBoost

This study proposes a multivariate dynamic cost standard prediction model based on random forest; LSTM is combined with XGBoost to solve the problem of accuracy in predicting complex cost standards in time series data. The model introduces random forest to sort the importance of multivariate features to screen out key features and improve the prediction performance of the model. Then the LSTM network is used to deeply learn the dynamic trend of the cost standard, which shows advantages in processing time series dependencies. Finally, the model is fine-tuned and predicted by the XGBoost algorithm to further improve the prediction accuracy of the model. The simulation results show that the model in this paper is significantly better than the traditional model in data prediction accuracy and can more effectively predict the dynamic changes of multivariate cost standards. By comparing the simulation results of different models, the integrated model based on RF-LSTM-XGBoost performs well in prediction accuracy, stability and data processing capabilities. This provides an efficient solution for dynamic cost standard prediction.