Research On The Basic Expenditure Analysis Model Of Financial Budget In Wulanchabu City

The year of 2016 is the first year of the 13th five-year plan.China's economy is in good shape,GDP grows by 6.7 percent in the whole year.The implementation of the structural reform of the national supply side,further improves the social productivity level,moderately expands the total demand,cuts overcapacity,de-stocking,de-leveraging,cost reduction,improves weak links,and becomes the main working point.From the production areas,governments strengthen the supply of high quality.At the same time,the local governments continue to intensify tax reductions and fees cut,full implementation of business tax to VAT policy,cancel some administrative controls and relaxation of supply constraints,These measures have a great influence on local fiscal revenues.At present,with the frequent use of the budget system,a large number of business data produced.Only as business using query and statements,the data's intrinsic value reflected not obviously,it can't provide scientific decision basis to ease the financial pressure for authorities,leading to a waste of resource.Therefore,this paper mainly based on ID3 decision tree and BP neural network model of budget analysis,summarizes the relationship between the historical data.Combining with the actual business,make analysis rules,through intelligent processing of the model,to reach the purpose of saving financial fund.Taking the past five years budget data of Ulanqab city departments as samples,using the data of the unit type,spending function subject,economic subject and spending amount information acquisition and processing,build two kinds of analysis model.From the aspect of ID3 decision tree,it is mainly to solve the problem of whether the new budget application data can be included in the budget arrangement.In the experiment,whether the sample data is included in the budget as the class label,in the budget is YSE,not in the budget is NO.After experimenting three times,the highest rate of the model accuracy was 90.6%,and the results were relatively good,but there was still room for improvement.From the perspective of BP neural network,it mainly solves the classification and judgment of the amount of the new expenditure.Classified the sample data,taking the range of each 500,000 Yuan as a class label,All of data are divided into 10 categories.According to the experiment,the highest rate of model accuracy is 84.8%,and there is a certain gap with the expectation.However,with the continuous expansion of business data and the continuous improvement of calculation methods,it is believed that the accuracy and rationality of the model can be further improved.Firstly,this paper briefly introduces the research contents,current situation and problems that need to be solved in the department budget of the city level.Secondly,explains the related technologies applied in the analysis process,including the decision tree ID3 model,BP neural network and MATLAB tools,etc.Finally,the paper elaborates the experiment process and concludes.