The application of GBDT integrated algorithm in the dynamic prediction of hospital financial distress
The traditional dynamic prediction model for financial distress has low accuracy when dealing with complex data in hospitals.To solve this problem,a study proposes using the minimum absolute contraction and selection operator algorithm to select financial indicators,and using the gradient enhancement decision tree ensemble algorithm to construct a dynamic prediction model for hospital financial distress.The results show that the highest accuracy of this model is 96%,with an F-value of 86%,and a G-value of 93%,all of which are higher than the other six dynamic financial distress prediction models used for comparison.And during the experiment,the research proposed that the running time and average error of the model were 23 seconds and 7%,respectively,which were lower than the comparison model.Experiments have shown that the dynamic prediction model for hospital financial distress based on gradient enhancement decision tree ensemble algorithm has more accurate and reliable prediction results,providing assurance for hospital financial management and improving the stability of hospital finance.
国家科技期刊平台
NSTL
万方数据
