Rick Anderson has categorised predatory publishing into four types:

Phony Journals. These are journals that falsely claim to offer documents based on legitimate and dispassionate scientific or scholarly inquiry. Such journals may be made available on either a toll-access or an open-access basis. One notable example from a few years ago was the Australasian Journal of Bone and Joint Medicine, which was published by Elsevier and presented to the world as a journal of objective scholarship, but was later revealed as a promotional sock-puppet for a pharmaceutical company—one such journal among several, as it turned out. Truly phony journals of this sort, though egregious, seem to be relatively rare.

Pseudo-scholarly Journals. These are journals that falsely claim to offer authors real and meaningful editorial services (e.g. peer review) and/or credible impact credentialing (e.g. Impact Factor), and thereby also falsely claim to offer readers rigorously vetted scientific or scholarly content. In this case, the content may or may not be legitimate scholarship—but the journal itself is only pretending to provide the traditional services of peer review and editorial oversight. This is perhaps the largest category of deceptive publisher, and also one of the more controversial ones, since the line between dishonesty and ineptitude can be fuzzy. For this reason, it makes sense to exercise caution in ascribing deceptive intent to these journals; however, in many cases (such as those that falsely claim to have an Impact Factor, that lie about their peer-review processes, or that falsely claim editorial board members), deceptive intent can be quite clear.

False-flag Journals. These are scam operators that set up websites designed to trick the unwary into believing that they are submitting their work to legitimate existing journals—sometimes by “hijacking” the exact title of the real journal, and sometimes by concocting a new title that varies from the legitimate one slightly.

Masqueraders. This looks like a variety of hijacking, except that there is no actual hijackee. In these cases, journals adopt titles designed to imply affiliation with a legitimate- and prestigious-sounding scholarly or scientific organisation that does not actually exist. For example, a Masquerader journal might call itself the American Medical Society Journal or Journal of the Royal Society of Physicians.

And a new form has emerged – those seeking to be republishers of works already published.

Andersen, 2015, Deceptive Publishing: Why We Need a Blacklist, and some suggestions on how to do it right, blogpost, The Scholarly Kitchen.