Table 9 phishGILLNET and competing methods characteristics
Method | Year of publication | Corpus (public/private/mix | Max data | Year when data source used generated | 3-class classification | Can handle unlabelled? |
|---|---|---|---|---|---|---|
Chan et al. [65] | 2004 | Public | 2.8 K | NA | No | Yes |
PILFER [10] | 2007 | Public | 7.8 K | 2002-2006 | No | No |
Abu-Nimeh et al. [11] | 2007 | Private | 2.8 K | 2005-2006 | No | No |
Bergholz et al. [16] | 2008 | Public | 8 K | 2004-2007 | No | No |
Abu-Nimeh et al. [12] | 2009 | Mix | 6.5 K | 2006-2007 | No | No |
Gansterer and Pölz [15] | 2009 | Mix | 15 K | 2007 | Yes | No |
Toolan and Carthy [19] | 2010 | Public | 8.3 K | 2004-2007 | No | No |
Bergholz et al. [17] | 2010 | Private | 40 K | 2007 | No | No |
Khonji et al. [20] | 2011 | Public | 8.2 K | 2003-2007 | No | No |
Al-Momani et al. [21] | 2011 | Public | 8.7 K | 2003-2007 | No | No |
phishGILLNET1 | - | Public | 8.7 K | 2003-2007 | No | No |
phishGILLNET2 | - | Public | 400 K | 2011 (phish-40 K, spam 320 K) | Yes | No |
2001 (good-40 K) | ||||||
phiahGILLNET3 | - | Public | 400 K | 2011 (phish-40 K, spam-320 K) | No | Yes |
2001 (good-40 K) |