Monitoring and Analysis of Venture Capital and Corporate Fraud Based on Deep Learning

Table 11

Winsorization change without eliminating missing value.

	(1)	(2)
	Probit	Poisson
	VC	−0.0483	−0.133
	(−2.01)	(−4.83)
Age	0.0133	0.0762
	(1.07)	(5.36)
Size	−0.0251	−0.0252
	(−2.54)	(−2.25)
ROA	−2.653	−2.152
	(−17.23)	(−27.42)
Dual	−0.118	−0.199
	(−5.05)	(−7.73)
Audit	−0.276	−0.446
	(−5.42)	(−6.76)
BdSize	0.0188	0.0289
	(0.35)	(0.48)
Tqmed	−0.0397	−0.0992
	(−1.07)	(−2.30)
Cash	−0.00302	−0.0364
	(−0.85)	(−4.27)
Constant	0.351	0.334
	(1.33)	(1.16)
Observations	20749	20800
Year and industry	Yes	Yes
Pseudo R^2	0.035	0.044
AIC	21517.6	36832.0
Log likelihood	−10666.8	−18316.0

p < 0.10, p < 0.05, and p < 0.01.