基于线性神经网络模型的政府财政支出结构预测——以新疆为例

Prediction on Government's Fiscal Expenditure Structure Based on Linear Neural Network Model:Case of Xinjiang

在分析1978—2010年新疆政府财政支出现状的基础上,运用线性神经网络模型,利用1995—2010年新疆政府财政支出的统计数据,对2011—2015年的相关数据进行预测,分析了此期间新疆政府财政支出的发展趋势。结果表明:2011—2015年期间,新疆政府的经济建设支出、基本公共服务支出、行政支出占财政支出总额的比重有所减小,国防支出可能呈平稳增长趋势——这与新疆所处的地理位置、自然环境、宗教文化等因素密切相关。最后结合新疆的特殊区情,提出了针对性的政策建议。

On the basis of analyzing the current situation of Xinjiang government's fiscal expenditure structure during 1978-2010,this paper uses the statistical data about Xinjiang government's fiscal expenditure from 1995 to 2010 to predict the future tendency of it during 2011-2015 by the linear neural network model.The result shows as follows:the proportions of economic construction,basic public service and administration in total fiscal expenditure would reduce during 2011-2015,but defense expenditure might rise continuously,which is closely related to geographic location,natural environment and religious culture of Xinjiang.Finally,it gives some suggestions according to the special condition of Xinjiang.