利用班福法则侦测您的数据中的异常.pdfVIP

SUGI 27 Statistics and Data Analysis Paper 249-27 Detecting Anomalies in Your Data Using Benford’s Law Curtis A. Smith, Defense Contract Audit Agency, La Mirada, CA ABSTRACT Analyzing large amounts of data looking for anomalies can be a disheartening task. You need techniques that will allow you to quickly assess the data in ways that will highlight potential anomalies while keeping you from chasing the wind. Benford’s Law is one such technique. Using Benford’s Law and the SAS© System you can quickly identify one or more first digit patterns in numeric variables that defy statistical averages. Within this paper, the author will present SAS code that will enable you to quickly and easily find anomalies in the data you analyze. The SAS code will include the Data Step, the Merge statement, and the FREQ, REPORT, and GPLOT procedures. The author will also present some findings from the data he analyzes. The technique presented is powerful, yet easy to understand and use. INTRODUCTION Benford’s Law is so named after Dr. Frank Law... I mean, Dr. Frank Benford. Dr. Benford was a physicist working for General Electric in the 1930s. He noticed that certain pages of his logarithm book were more worn than others. After some study, he r