[01:38:02] *** Joins: jstar- (~jstar@77.235.50.85)
[01:39:23] *** Quits: jstar (~jstar@unaffiliated/jstar) (Ping timeout: 248 seconds)
[01:39:23] *** jstar- is now known as jstar
[01:39:24] *** Quits: Flammie (~flammie@80.111.32.7) (Ping timeout: 248 seconds)
[01:41:27] *** Quits: crazydiamond (~crazydiam@178.141.72.105) (Remote host closed the connection)
[04:11:15] *** Quits: drbean (~drbean@124.219.82.83) (Ping timeout: 265 seconds)
[04:15:49] *** Joins: drbean (~drbean@124.219.83.158)
[04:26:03] *** Quits: drbean (~drbean@124.219.83.158) (Ping timeout: 256 seconds)
[04:30:34] *** Joins: drbean (~drbean@124.219.82.219)
[04:38:45] *** Quits: drbean (~drbean@124.219.82.219) (Ping timeout: 248 seconds)
[05:04:24] *** Joins: drbean (~drbean@124.219.82.135)
[09:41:58] *** Joins: vinit-ivar (~vinit@122.170.50.86)
[09:44:10] *** Quits: vin-ivar (~vinit@122.169.13.180) (Ping timeout: 244 seconds)
[11:10:34] <inariksit> vinit-ivar: did you get any results from the experiment?
[12:13:28] *** Joins: Main (~main@dhcp-065065.wlan.ntnu.no)
[12:14:01] *** Parts: Main (~main@dhcp-065065.wlan.ntnu.no) ()
[13:47:37] *** vinit-ivar is now known as vin-ivar
[13:47:40] <vin-ivar> hey inariksit
[13:47:51] <vin-ivar> i haven't done much after formatting the data, I had an exam today
[13:47:58] <vin-ivar> will play around with it tonight
[13:55:48] <inariksit> cool :)
[13:57:31] <vin-ivar> i wonder if it'll work, it ought to, logically
[14:57:04] *** Joins: BobL (82f1310b@gateway/web/freenode/ip.130.241.49.11)
[15:34:52] *** Quits: BobL (82f1310b@gateway/web/freenode/ip.130.241.49.11) (Ping timeout: 246 seconds)
[16:05:45] *** Joins: BobL (82f1310b@gateway/web/freenode/ip.130.241.49.11)
[16:53:26] *** Flammie_ is now known as Flammie
[18:48:33] *** Joins: jmvanel (~jmvanel@78.193.21.40)
[19:42:09] *** Quits: jmvanel (~jmvanel@78.193.21.40) (Quit: Quitte)
[20:30:27] <vin-ivar> inariksit:
[20:30:33] <vin-ivar> some good news
[20:30:55] <vin-ivar> i wrote a basic classifier, decided to see whether it could estimate whether <vblex> belongs or not
[20:31:02] <vin-ivar> given the GF tags
[20:31:09] <vin-ivar> the accuracy is around 0.89 :)
[20:31:23] <vin-ivar> so now I'll try to convert it to a multilabel classifier
[20:34:28] *** Quits: BobL (82f1310b@gateway/web/freenode/ip.130.241.49.11) (Ping timeout: 246 seconds)
[21:13:53] <vin-ivar> okay, the lowest accuracies are for <p2> and <p3> (second/third person)
[21:14:00] <vin-ivar> 0.73 and 0.72
[21:26:15] *** Joins: crazydiamond (~crazydiam@178.141.72.248)
[21:56:24] *** Quits: vin-ivar (~vinit@122.170.50.86) (Read error: No route to host)
[21:58:40] *** Joins: vin-ivar (~vinit@122.170.50.86)
[22:46:22] <vin-ivar> okay, so the problem with assigning labels right now is that it's pretty trigger happy with labels
[22:46:34] <vin-ivar> i'll work out something that also considers already assigned labels
[22:48:05] <vin-ivar> btw, https://github.com/vinit-ivar/apertium-gf/commit/85edc15cc5e21e6a1906157fafc69327bd0e8a8e