There are about 10^10 Internet users, so if we think it's very unlikely that it's her (10^-6 chance compared to anyone), that gives a 10^-16 prior chance of bei

There are about 10^10 Internet users, so if we think it's very unlikely that it's her (10^-6 chance compared to anyone), that gives a 10^-16 prior chance of being her. Assuming that they post almost every day and miss a day every few months, the probability of missing a single day for a random user is probably around 10^-2, so if the account didn't post during her public activities for around 13 times, that gives a 10^26 Bayes factor and thus we can conclude it's her or someone intentionally trying to frame her with 1 - 10^-10 > 6 sigma equivalent certainty. Obviously a more precise analysis is possible.