EURASIP Journal on Information Security

Table 2 phishGILLNET1--PLSA word/topic probability distribution

From: phishGILLNET—phishing detection methodology using probabilistic latent semantic analysis, AdaBoost, and co-training

Topic (z) (phishing)		Topic (z) (non-phishing)
Word ( w )	Probability P ( w \| z )	Word ( w )	Probability P ( w \| z )
Bank	0.058	Ocean	0.024
Online	0.046	Honolulu	0.014
Banking	0.033	Imminent	0.013
America	0.032	Assuring	0.010
Account	0.021	Handsome	0.009
Update	0.019	Builder	0.007
Security	0.017	Lush	0.005
Customer	0.014	Lousy	0.005
Below	0.013	Roads	0.005
Link	0.013	Vantage	0.005
Click	0.011	Sweetness	0.005
Please	0.011	Wine	0.004

Back to article page