Bayes Classifying Analysis of Earnings Manipulation in Listed Companies in China

LI Yan-xi,GAO Rui,BAO Shi-ze,YAO Hong
DOI: https://doi.org/10.3969/j.issn.1003-5192.2007.03.010
2007-01-01
Forecasting
Abstract:Correctly recognizing the earnings manipulating behaviors of listed companies is very important for the healthy development of stock market in China.Based on the analyses of motivations,main tricks and appearances of aggressive earnings manipulation,we initially design 14 financial indexes which hold certain matching relationships and interrelationships among each others.Then,after further screening,we finally select 5 indexes from the initial 14 indexes.We select 35 listed companies which have been punished by China Securities Regulatory Commission as the earnings manipulation samples;and also select 35 healthy listed companies with the same firm size and same industry as the 35 punished companies as the control samples.Through using Bayes classifying analysis to compare the two groups of samples,we establish a classification model of earnings manipulation.The empirical results indicate that the model can obtain good classification effects.
What problem does this paper attempt to address?