Open in new window / Try shogun cloud
--- Log opened Sat May 14 00:00:58 2011
-!- blackburn [~qdrgsm@188.168.3.72] has quit [Quit: Leaving.]01:44
-!- alesis-novik [~alesis@188.74.87.84] has quit [Quit: I'll be Bach]03:41
-!- serialhex [~quassel@99-101-148-183.lightspeed.wepbfl.sbcglobal.net] has quit [Remote host closed the connection]03:47
-!- serialhex [~quassel@99-101-148-183.lightspeed.wepbfl.sbcglobal.net] has joined #shogun04:27
-!- serialhex [~quassel@99-101-148-183.lightspeed.wepbfl.sbcglobal.net] has quit [Remote host closed the connection]05:01
-!- alesis-novik [~alesis@188.74.87.84] has joined #shogun07:10
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has quit [Ping timeout: 264 seconds]13:51
-!- serialhex [~quassel@99-101-148-183.lightspeed.wepbfl.sbcglobal.net] has joined #shogun15:01
-!- sudarshan [~sudarshan@c-174-50-225-74.hsd1.tn.comcast.net] has joined #shogun15:28
-!- sudarshan [~sudarshan@c-174-50-225-74.hsd1.tn.comcast.net] has quit [Remote host closed the connection]15:36
-!- blackburn [~qdrgsm@188.122.225.203] has joined #shogun15:48
-!- serialhex [~quassel@99-101-148-183.lightspeed.wepbfl.sbcglobal.net] has quit [Remote host closed the connection]17:46
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has joined #shogun21:04
CIA-31shogun: Soeren Sonnenburg master * rb9d67bb / (5 files in 4 dirs): add SGMatrix support in python typemaps - http://bit.ly/jB5H5I22:36
@sonney2kblackburn, I only now realized that we have another naming clash22:41
@sonney2kwhat I call sparse matrix now is actually only a sparse vector22:41
@sonney2kso I will have to do renamings to SGSparseVector etc22:42
@sonney2kand introduce SGSparseMatrix22:42
@sonney2ksame for strings btw22:42
@sonney2kSGString22:42
@sonney2kwill need to introduce SGStringList22:43
blackburnsonney2k: ehh, but do we have sparse vectors already?22:44
@sonney2ksure but called SGSparseFeatureMatrix or so22:45
blackburnah22:45
blackburnsonney2k: I need some advice with LLE22:46
blackburnwhat is better - to allocate memory for all instance life or only within init()?22:46
@sonney2kblackburn, ask but I am not sure if mlsec/cwidmer could help better here22:47
@sonney2kblackburn, I don't understand22:47
blackburnsonney2k: not about algorithmic but about pure programming ;)22:47
@bettyboorahaha!?!22:47
@sonney2kok22:47
blackburnI mean I have to allocate some matrices22:47
blackburnshould I allocate it within init() method22:47
blackburnor better allocate it with constructor?22:47
@sonney2kblackburn, do you know the sizes before calling init?22:48
blackburn:D22:49
@sonney2kdo you use them 'often'22:49
blackburnsorry of course not22:49
blackburnoh how stupid I am22:49
@sonney2kok22:49
blackburnsonney2k: another question about that preproc framework22:50
blackburnehh.22:50
blackburnwell, we have apply_to_feature_matrix and apply_to_feature_vector22:51
blackburnif to matrix I understand, it will call init, etc22:51
blackburnbut what about vector?22:51
blackburnshould I use already initialized features?22:51
blackburnsonney2k: https://github.com/SergeyLisitsyn/shogun/blob/274a2fefe4de3c82dc5c585e89ab3edd336a12b0/src/libshogun/preproc/LLE.cpp22:52
@sonney2kblackburn, you have to assume that your preproc is inited22:52
blackburnnow all routines are in init22:52
blackburnI'm stucked what to do with apply_* methods22:53
@sonney2kso that you can apply it to single vectors (when the feature matrix is not avalable in memory)22:53
@sonney2kblackburn, btw we have a CMath::display_matrix()22:54
blackburnsonney2k: ok, will know22:54
blackburnit just some crap not cleaned up22:54
blackburnof course there wouldn't be any sg_print :)22:54
blackburnsonney2k: so, when apply_to_feature_matrix is called I should re-init (or init) with given features, right?22:55
@sonney2kblackburn, never re-init22:56
@sonney2kalways assume that you are inited22:56
blackburnI'm confused what exactly I should do in init()22:56
blackburnsonney2k: if it is inited already why CFeatures instance is passed to apply_to_feature_matrix?22:59
@sonney2kblackburn, I don't understand - it should apply ti to the CFeatures object that is passed...23:01
blackburnbut what should init(CFeatures* f) do?23:01
@sonney2kexample:23:02
blackburnI can't see any relation between two different features..23:02
@sonney2kyou want to remove features which have small variance23:02
@sonney2kthen you measure variance on init() and recall which dims should be removed23:02
@sonney2kthen when you call apply* you use that list23:03
blackburnah!23:03
blackburnsonney2k: but what about dim reduction techniques like LLE? it is absolutely determined to feature set23:05
blackburnwe can't init on one feature set and process other23:05
blackburndo you understand my miscomprehension?23:06
@sonney2kno23:07
blackburneeeh23:07
blackburnnevermind :)23:07
@bettyboocool23:07
blackburnwill try to understand how it is done with pca23:07
blackburnah don't understand how to do it with LLE23:14
blackburnsonney2k: will you implement addition of two SGVectors?23:20
blackburnand I'm interested with scalar multiplication too23:20
CIA-31shogun: Soeren Sonnenburg master * r7b958bf / (57 files in 14 dirs):23:24
CIA-31shogun: rename sparse matrix to sparse vector and introduce sparse matrix23:24
CIA-31shogun: structure and string list structure - http://bit.ly/j1Sl1123:24
@sonney2kblackburn, I won't - but feel free to do so23:25
@sonney2kbtw, have a look at DataType.h and tell me if that is OK23:25
blackburnI don't understand why you are asking me for that (in fact I can't grasp preproc) :)23:26
blackburnoh nice I switched branch and lost LLE :D23:29
@sonney2khard to believe23:29
@sonney2kswitch back!23:29
blackburnaha23:29
blackburnwill get from github23:29
blackburnbut I was surprised23:30
blackburnEHRRR i'm useless23:31
@sonney2kblackburn, bah get some sleep or eat more chocolate ;-)23:32
@bettyboo8)23:32
blackburnsonney2k: should ops on vectors be implemented as its methods?23:34
@sonney2kblackburn, yes makes most sense - at some point this does no longer belong in datatype.h though - but for now yes23:35
blackburnsonney2k: SGVector.add(a,b)?23:36
@sonney2kno only add(v)23:36
blackburnit is less readable, don't you think so/23:36
blackburn?23:36
blackburnah23:38
blackburnI see23:38
@sonney2ki think it is more readable - it adds v to the current object23:44
blackburnmoreover it doesn't copy vector23:45
blackburnbut what if we want x,y and x+y both?23:45
@sonney2k sorry what?23:47
@sonney2kahh I understand23:47
@sonney2kwell then feel free to add another function for this23:48
blackburnsonney2k: what about zero vector constructor?23:49
@sonney2kblackburn, implement what you need...23:50
blackburnok.23:53
@sonney2kblackburn, you probably can take much of what is in CMath* now23:54
@sonney2ke.g. dot, fill vector range_fill vector etct etc23:55
blackburnanyway my addition doesn't compile yet :D23:56
-!- warpy [~warpy@bzq-79-183-44-64.red.bezeqint.net] has quit [Ping timeout: 252 seconds]23:56
--- Log closed Sun May 15 00:00:00 2011