基于Python计算生态的财务异常数据检测

Abnormal Financial Data Detection Based on Python Computing Ecology

财务信息系统由于各种原因会包含一些异常记录,对这些异常记录的检测,在财务审计中有着重要的意义。文章首先分析了判别财务数据异常的统计参数指标,然后总结了几种常用的检测方法,最后在Python计算生态中的numpy、pandas、matplotlib、scikit-learn工具包所提供的统计运算、可视化功能、数据挖掘和机器学习的功能基础上,设计了一个基于Python计算生态挖掘财务异常数据的原型框架。

The financial information system contains some abnormal records due to various reasons,and the detection of these abnormal records is of great significance in financial auditing.This paper first analyzes the statistical parameter indicators for judging abnormal financial data,then summarizes several commonly used detection methods,and finally designs a prototype framework for mining abnormal financial data based on Python computing ecology through statistical operations,visualization functions,data mining and machine learning provided by numpy,pandas,matplotlib,scikit-learn tool packages in Python computing ecology.