Open in new window / Try shogun cloud
--- Log opened Fri Feb 06 00:00:07 2015
-!- nszceta [~nszceta@] has joined #shogun00:33
AndreLopesIn this example01:26
AndreLopesI got a warning01:26
AndreLopes[WARN] In file /builddir/build/BUILD/shogun-96f3cf3ce58514299f98e688b7c43e057ad4fa41/src/shogun/classifier/AveragedPerceptron.cpp line 90: Averaged Perceptron algorithm did not converge after 10 iterations.01:28
AndreLopesIs this normal to the example ?01:28
AndreLopesAnd why does the neural network only accepts a DynamicObjectArray ?01:44
AndreLopesAnd also this.. SwigType_p_p_shogun_CSGObject01:44
-!- nszceta [~nszceta@] has quit [Quit: Textual IRC Client:]02:21
-!- PirosB3 [] has quit [Quit: PirosB3]02:47
-!- AndreLopes [bae787ab@gateway/web/cgi-irc/] has quit [Quit: - A hand crafted IRC client]03:54
shogun-buildbotbuild #962 of nightly_default is complete: Failure [failed notebooks]  Build details are at  blamelist: sanuj <>, Kunal Arora <>, Fernando Iglesias <>, Bj?rn Esser <>04:08
-!- lambday [c40f1064@gateway/web/freenode/ip.] has joined #shogun07:02
-!- mode/#shogun [+o lambday] by ChanServ07:02
-!- chintak [6acef958@gateway/web/cgi-irc/] has joined #shogun07:36
-!- chintak [6acef958@gateway/web/cgi-irc/] has quit [Quit: - A hand crafted IRC client]08:21
-!- besser82 [~besser82@2a02:8108:8800:3d00:f2de:f1ff:fe89:42d4] has joined #shogun09:22
-!- besser82 [~besser82@2a02:8108:8800:3d00:f2de:f1ff:fe89:42d4] has quit [Changing host]09:22
-!- besser82 [~besser82@fedora/besser82] has joined #shogun09:22
-!- mode/#shogun [+o besser82] by ChanServ09:22
-!- sonne|work [] has quit [Quit: Leaving.]11:33
@besser82lisitsyn, ping? ^^12:09
@besser82lisitsyn, can you please give me commit-access to your swig-fork on github?  =)12:10
-!- lambday [c40f1064@gateway/web/freenode/ip.] has quit [Ping timeout: 246 seconds]12:38
@lisitsynbesser82: oops sorry13:13
@lisitsynyeah sure13:13
@lisitsynbesser82: have fun :)13:15
@besser82lisitsyn, thx ^^13:15
-!- Netsplit *.net <-> *.split quits: aditya9513:19
-!- Netsplit over, joins: aditya9513:21
@besser82lisitsyn, force pushed rebased commit into bugfix-branch13:28
@besser82lisitsyn, you should be able to resolve that by:  git checkout bugfix/apple_clang_check && git fetch --all && git reset --hard origin/bugfix/apple_clang_check13:29
-!- exploreshaifali [~shaifali@] has joined #shogun13:38
-!- PirosB3 [] has joined #shogun14:50
-!- exploreshaifali [~shaifali@] has quit [Ping timeout: 256 seconds]15:28
-!- exploreshaifali [~shaifali@] has joined #shogun16:06
-!- sonne|work [] has joined #shogun16:18
sonne|worklisitsyn: around?16:19
@lisitsynsonne|work: yes16:21
sonne|worklisitsyn: I have a ML kind of problem - I have a few distributions (say 10)16:21
sonne|workand some mixture of these 10 distributions with unknown coefficients16:22
sonne|workI am now looking for a way to estimate the coefficients16:22
sonne|workand these given distributions are just based on counts - so histograms16:23
@lisitsynsonne|work: any assumption on distributions? :)16:24
@lisitsynwhy not to fit GMM?16:25
sonne|workI could smooth them all :)16:25
sonne|workwell the base distribs are not gaussians16:25
@lisitsynthen choose something general and fit it with l-bfgs maybe?16:25
sonne|workor some gradient decent kind of thing ... hmmhh sure would work16:26
@lisitsynsonne|work: scipy has cool fmin_bfgs just give it a task16:27
sonne|worklisitsyn: yeah that would do it16:28
@lisitsynsonne|work: can you assume gamma distribution16:29
sonne|worklisitsyn: no the histograms can be anything even some multimodal stuff16:29
sonne|worklisitsyn: but that's good enough16:30
@lisitsynsonne|work: fitting multimodal mixture sounds scary :D16:30
sonne|worklisitsyn: relax all 1d - not so scary16:31
@lisitsynwhy not to assume # of distributions equals # of modes?16:31
@lisitsynthen gaussian is good16:32
sonne|workall I know is that the resulting distribution is a mixture of the others - and I would 'just' want to know the most likely mixture coefficients16:35
sonne|workif the match is good enough then I can infer voila the distribution did get generated from the following sources16:35
-!- abinashm [cb6ef316@gateway/web/freenode/ip.] has joined #shogun16:47
-!- abinashm [cb6ef316@gateway/web/freenode/ip.] has quit [Quit: Page closed]17:28
-!- hargup [uid10222@gateway/web/] has joined #shogun18:12
-!- abinashm [abinashmeh@gateway/shell/waartaa/x-giqmdrqtnbgvniot] has joined #shogun18:18
-!- abinashm [abinashmeh@gateway/shell/waartaa/x-giqmdrqtnbgvniot] has left #shogun []18:20
-!- abinashm_ [uid66179@gateway/web/] has joined #shogun18:23
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 252 seconds]19:48
@wikinganybody here?21:14
-!- PirosB3 [] has quit [Quit: PirosB3]21:15
@lisitsynwiking: ja21:16
@wikinglisitsyn: how do u use numpy array as a RealFeatures?21:17
@wikingbtw pytthon modular is totally broken still on osx21:17
@lisitsynwiking: yeah I know21:17
-!- sonne|osx [] has joined #shogun21:30
@wikinglisitsyn: here?22:10
@wikinglisitsyn: how the hell do u matplotlib a 2d numpy array, where each element is [a,b]. and the whole numpy array is like [[a,b],[x,v]...]22:11
@lisitsynwiking: hmm22:24
@lisitsynwiking: dont' get it22:24
@lisitsynyou need 2xN matrix?22:24
@wikingi just want to plot a 2d graph22:27
@wikingwhere the input is22:27
@wikingimean [[a,b],[x,v]...]22:28
@wikingbut if i do axis.plot(x...)22:28
@wikingthen i just get one part of each element of the arrays (the first) as y and x is basicaly the lenght of the array22:28
-!- exploreshaifali [~shaifali@] has quit [Quit: Leaving]22:36
--- Log closed Sat Feb 07 00:00:09 2015