TrustFlow results
Nov. 3rd, 2006 16:47
i checked out TrustFlow II for LiveJournal. it measures which people not on my flist are "close" by virtue on being on the lists of my friends and so on down the line.
michaelsullivan,
puffinboy,
charlottezweb (400 - 450)
matthewsimon,
la_penguinita,
henrygrosvenor,
tranquilisea,
ana (450 - 500)
it,
berkeleyfarm,
eleanorjevans,
owned_by_2_cats (500 - 550)
selki,
royceedward,
lilairen,
lynn_massa (550 - 600)
enna_aral,
srunning,
stevendj,
okoshun,
marykaykare (600 - 650)
micheinnz,
womzilla (650 - 700)
rosefox,
pegkerr (700 - 750)
misia,
mactavish,
rmjwell,
crowleycrow,
masaiwarrior,
skzbrust,
brisingamen (750 - 800)
curiousangel,
pperiwinkle,
therealjae,
annburlingham (800 - 850)
ellen_kushner,
trinker,
wandra,
sartorias,
baratron (850 - 900)
mrissa,
eeyorerin,
patgreene (900 - 950)
eleanor,
opportunitygrrl,
leback,
annafdd (950 - 1000)
dirtyanswer,
noctilucentious,
nudas_veritas,
pogo_girl,
lost_cities,
annaoj,
carbonel (1000 - 1050)
sdn,
cattitude,
nancylebov,
sinboy,
australian_joe,
1crowdedhour,
rachelmanija (1050 - 1100)
gerisullivan,
wispfox,
lisajulie (1100 - 1150)
brian1789,
griffen,
spiritrover,
minnehaha,
johnpalmer,
diony (1150 - 1200)
bohemiancoast,
fivemack,
rushthatspeaks,
barbarakitten_t,
lissamc,
solcita (1200 - 1250)
browngirl,
nooks,
laurel,
pfunque,
hkbear,
rozk,
daveyfunsize,
elynne_smut,
bitty,
daystreet (1250 - 1300)
the_ogre,
dduane,
peake,
hopeforyou,
fantasygoat,
sevoo,
oursin (1300 - 1350)
beamjockey,
erin,
filkerdave,
angilong,
copperwise,
kijjohnson (1350 - 1400)
interesting algorithm; it seems to privilege small networks over large ones, but looks in general like i would expect (people whom i actually know but have purposely not flisted are good test cases, as are people whom i know but whose LJ's i don't read because they don't post :). there are very few names i don't recognize at all; i'll check them out when i run into one of those periods where i actually have more energy to read LJ. and hey, a couple of secondary journals i didn't know about.
via
machineplay.
no subject
on 2006-11-04 04:50 (UTC)trustflow II
on 2006-11-04 06:42 (UTC)i wouldn't translate the metric as "most" or "least" trusted. it's really not trust as people use the term normally; it's trust in a networking sense, which is an abstracted way to look at it. frex, google page ranking algorithm is a trust metric, as is ebay's feedback, and slashdot's karma.
the terminology always confuses people, which is why i used "close" as the term instead of "trust" -- the smaller numbers are journals closer to mine in LJ space; more people who are more closely connected to me have them flisted than the ones with larger numbers. the numbers themselves are units internal to the algorithm used.
Re: trustflow II
on 2006-11-04 17:13 (UTC)no subject
on 2006-11-04 07:12 (UTC)no subject
on 2006-11-06 15:50 (UTC)