基于数据挖掘技术的温州市财政收入的影响因素分析及预测

Analysis and Prediction of Influencing Factors of Fiscal Revenue in Wenzhou Based on Data Mining Technology

摘要: 本文首先建立了Lasso变量选择模型，挑选出了影响温州市财政收入的四个关键性因素，即在岗职工工资总额、城镇居民人均消费性支出、地区生产总值和税收。在此基础上，分别针对这四个影响因素建立GM (1, 1)模型，得到它们2016~2018年的预测值。最后，我们通过训练合适的BP神经网络模型，给出了温州市财政收入在2016~2018年的预测值。

Abstract: This paper established a Lasso model, which selects four key factors affecting the financial reve-nue of Wenzhou, such as the total wages of staff and workers, per capita consumption expenditure of urban residents, GDP and tax. On this basis, we also get the forecast values of above four factors especially from 2016 to 2018 by using GM (1, 1) model. Finally, we give the forecast values of financial revenue of Wenzhou from 2016 to 2018 by building an appropriate BP neural network model.