修正Benford分布律及其模拟研究

Research of Corrected Benford's Distribution and Its Statistical Simulation

Benford分布律是常用的数据质量评估方法。通常,Benford分布律只适用于完整数据集的数据质量评估。对于完整数据集的有界子集,提出修正Benford分布律评估其数据质量,拓宽了Benford分布律的适用范围。随机模拟结果显示,新方法的统计性质比Benford分布律更好,评估结果更合理。

Special Data Dissemination Standards（SDDS） of IMF has been accepted in China, it means that there is a challenging goal for studying and perfect evaluation methods of data quality. Benford＇s distribution has been widely applied in the data accuracy assessment. In General,Benford？s distribution can only be applied to complete recorded data quality assessment. This paper enhance Benford＇s distribution with the corrected Benford？s distribution to handle situations where data record is bounded. The simulation studies show that improved statistical property over traditional Benford＇s distribution in data quality assessment.