基于RF-GM(1,1)-BP模型预测福州市财政收入

Predicting Fuzhou City’s Fiscal Revenue Based on RF-GM(1,1)-BP Model

摘要: 掌握市场趋势和规划收支费用对于财政部门而言具有极为重要的意义。本文选取1994~2021年福州市年财政收入相关数据，采用随机森林(RF)模型识别出财政收入的关键影响特征，随后建立GM(1,1)-BP组合模型，对2022~2025年福州市的年财政收入进行预测。预测结果表明RF-GM(1,1)-BP组合模型十分适合用于预测福州市财政收入；同时，福州市财政收入将在2021年之后稳步增长，并在2025年到达7,999,256万元。该结论能为相关部门实施的决策提供一定的理论参考。

Abstract: Mastering market trends and planning revenue and expenditure expenses is of great significance for the financial department. This paper selects the relevant data of Fuzhou’s annual financial rev-enue from 1994 to 2021, uses the random forest (RF) model to identify the key impact characteris-tics of financial revenue, and then establishes a GM(1,1)-BP combination model to predict the annu-al financial revenue of Fuzhou from 2022 to 2025. The prediction results indicate that the RF-GM(1,1)-BP combination model is very suitable for predicting the fiscal revenue of Fuzhou City; Meanwhile, the fiscal revenue of Fuzhou City will steadily increase after 2021 and reach 79992.56 million yuan by 2025. This conclusion can provide a certain theoretical reference for decision- making implemented by relevant departments.