Quality Assessment Of The Statistical Data In Guilin Using Benford Laws And The Random Number Assistant Survey Method

The statistical data is the basis of statistical work.No quality data is a bunch of useless numbers,and its existence is still very harmful.So the quality of statistical data is the key to the statistical work in the whole process.Therefor,it is very important to find out one or more suitable methods to improve the quality of statistical data.The quality of statistical data is mainly reflected in four aspects:accuracy,timeliness,comprehensiveness and systematicness.Accuracy is the core of the quality of statistical data,to engage in quasi statistical data is the bounden duty of statisticians,but also the first task of the statistical department.In this paper,we will use two different methods to evaluate the accuracy of the statistical data of Guilin from the two aspects of"seeking truth" and "checking truth".In the "checking truth" aspect of the data,quality detection.In this paper,Benford rule is used to test the four indicators of the data of the residents' living standard of Guilin Bureau of Statistics.Using the Excel software to the first and second digits of the sample data for statistical detection and analysis,the results show that the distribution law of the first digit and the second digit is well fitted with the Benford rule.After that,the correlation coefficient,goodness of fit,K-S detection and distance detection are used to evaluate the overall number.The results show that the numerical distribution of the sample data is well fitted with the Benford rule,so the data to be detected is accurate and reliable.In turn,it also shows that the Benford rule has a good applicability to the quality of the statistical data of Guilin.In the "seeking truth" aspect of the data,the data quality control.This paper introduces a new data "truth-seeking" method,namely,the random number assistant survey method,which is based on the law of large numbers and gives the concrete theoretical proof.Then,in terms of its specific application,taking the F value generated by the random number aided survey method in four industry in 2013,using SPSS software for normality test,and then carrying out statistical analysis,confirmed the method is very feasible.After that,the paper puts forward some suggestions combining with the adjustment of F value,the optimization of the specific operation process and various auxiliary investigation methods.Finally,the paper makes a summary and outlook,pointing out the shortcomings of the application of this paper,and hopes to continue to study and supplement,in order to maximize the use of these two methods to assess the quality of Guilin statistical data.