Well done, it found my ancient old account.
Show HN: Using stylometry to find HN users with alternate accounts
31–40 of 562 posts
Re: Show HN: Using stylometry to find HN users with alternate accounts
#32Amusingly can't run it on the author since not enough comments
Re: Show HN: Using stylometry to find HN users with alternate accounts
#33Not today. You fail, I win.
Nice. Just out of curiosity are you taking any countermeasures or varying your writing style across accounts in any way?
Re: Show HN: Using stylometry to find HN users with alternate accounts
#34My nearest match is only at 0.406. It'd be interesting to see who the most unique commenters are, but it's also quite possible it wouldn't be flattering.
Re: Show HN: Using stylometry to find HN users with alternate accounts
#35Wow... how !
Re: Show HN: Using stylometry to find HN users with alternate accounts
#36Earlier quoted context omitted.
If it's >0.9 is you can almost guarantee it's an alt but I've seen certain matches at 0.6. The problem is writing styles change over time. Another idea I had was converting the scores which are just cosine similarity scores into percentiles (so 0.99 would be 99th percentile of certainty) to make them more human interpretable.
I make new accounts every so often and the accounts of mine that it found have a score of around 0.3. I'm not actively trying to defeat stylometry but it's possible I just have a particularly unremarkable writing style.
Well I must be stereotypical myself because it found me at 0.8 !
Re: Show HN: Using stylometry to find HN users with alternate accounts
#37It would be nice to make the names clickable. I don't think the list of pg alternate account is accurate. I checked a few. They have many oneliners that is typical of pg, but the topics and style don't look similar. I searched a few more and got better results. :) I searched myself (that I know that I have no alternate accounts). I recognize a few users that are interested in similar topics, and I discuss/upvote them…
> I searched myself (that I know that I have no alternate accounts). I recognize a few users that are interested in similar topics, and I discuss/upvote them many times. But I didn't recognize most of the user of the list.
It's based purely off frequency of the 200 most common English 1 word phrases, 2 word phrases, 3 word phrases, 1 character sequences, 2 character sequences, and 3 character sequences. Topic does not really have anything to do with it. If I had more time I probably would've done a smarter model that accounted for things like that.
Re: Show HN: Using stylometry to find HN users with alternate accounts
#38Hmmm, doesn't seem to work. But you have convinced me (and many others?) to search our alts consecutively and so now do know who has alts?
Re: Show HN: Using stylometry to find HN users with alternate accounts
#39totally on spot
my current and my old account
Re: Show HN: Using stylometry to find HN users with alternate accounts
#40Sticking myself in (I haven't ever had another account) my closest match (at 0.43) is the maintainer of an Open Source project which I have occasionally commented about. They are also British, as am I.
My guess is that as they commonly mention the project and I have on a number of occasions, that has formed the link. Plus maybe usage of common British terms, but that seems far less significant.
It's super interesting!
It would be good if there were more controls to filter the type of words and language that are used for the matching algorithm. So you could say exclude words not in the dictionary. I wander how that would effect my link with this other person.