We have no idea what the true rate of false accusations is. We know that the number where the woman is PROVED to be lying is low, but there is an unknown number where it is false, it just can't be proved.
If person A accuses B of rape, and it is then found out that A never even met B and was elsewhere at the time of the alleged rape, then the case is a provable false accusation. That MAY (depending on what happens as a result) make it into false accusation statistics.
If however, person A accuses B of rape, but actually the sex was consensual (though maybe regretted), B may be found not guilty of the rape, but its impossible to prove that A made a false allegation, unless she outright admits it. So it wouldn't make it in to false accusation statistics.
Given the vast majority of rape cases turn on the issue of consent, where false accusations can rarely be proved, it is impossible to know what the true rate of them is.
Coupled with that, some women (or men of course) may make an allegation of rape that they feel is completely founded, but that isn't actually rape. ie where someone feels 'I don't know if this was rape, but I feel it was wrong' following crappy sexual behaviour from someone, or equally, where someone simply regrets the sex the next day and over time they start to wonder if they truly consented after all, because they wouldn't have done so when sober.
None of those would make it into the 'false allegation' statistics.
The ones that do are in the tiny minority, so they actually mean very little.
Sex offences are also treated differently in that the complainant gets anonymity, whereas with a violent offence, they will usually not.
I really think the time has come to change the law on this, and give sex suspects the right to anonymity.