--- Log opened Wed Jun 29 00:00:29 2011 | ||
-!- f-x [~user@117.192.197.13] has joined #shogun | 00:02 | |
blackburn | whoaa | 00:03 |
---|---|---|
@sonney2k | f-x, around? | 00:17 |
@sonney2k | if so what do you think - did my comment make sense | 00:17 |
@sonney2k | ? | 00:17 |
f-x | sonney2k: yeah, present | 00:19 |
f-x | sonney2k: actually just commented on github on that very comment | 00:19 |
f-x | it does make sense, but what would go into say a CSimpleStream class? | 00:20 |
f-x | only the pure virtual get_vector() function, i'm thinking | 00:20 |
@sonney2k | yes the interface | 00:21 |
f-x | okay, so at least i understood correctly | 00:22 |
@sonney2k | so you could rely on these functions being available and don't have to do extra checks | 00:22 |
f-x | that's a nice thing to have | 00:22 |
f-x | and can i use templating? | 00:22 |
f-x | for the CSimpleStream class? | 00:22 |
@sonney2k | f-x, you run into the inheritance problem then | 00:23 |
@sonney2k | so no | 00:23 |
@sonney2k | you could however use overloading if that doesn't break | 00:23 |
f-x | sonney2k: are you sure about that inheritance problem? | 00:24 |
@sonney2k | yes | 00:24 |
blackburn | hmmm how to shorten Distance-based Dimension Reduction Preprocessor | 00:24 |
blackburn | haha base could be class NLDR (nonlinear dimension reduction) and derived is dbdrplandmarkmds | 00:25 |
blackburn | sonney2k: you should like it - deutsch have looong words too :) | 00:26 |
@sonney2k | blackburn, I have no idea | 00:26 |
blackburn | sonney2k: just joking ;) | 00:26 |
blackburn | sonney2k: is it a complete C.elegans genome in our data? | 00:27 |
blackburn | how bad we can't run this dna :( | 00:28 |
blackburn | I would want to see a huge worm with legs or so | 00:28 |
f-x | sonney2k: how about templating just the derived CSimpleStreamFromSimpleFeatures and CSimpleAsciiStream? | 00:30 |
f-x | i think that could work | 00:30 |
@sonney2k | f-x, I don't understand what exactly you want to do | 00:35 |
@sonney2k | then you have no read baseclass - right? | 00:36 |
@sonney2k | err SimpleStream | 00:36 |
f-x | sonney2k: no, all that is still there | 00:36 |
f-x | it changes one thing | 00:36 |
@sonney2k | f-x, but no interfaces or? | 00:36 |
f-x | in the StreamingSimpleFeatures class | 00:37 |
f-x | set_vector_reader just makes the reader a generic "get_vector" function | 00:37 |
f-x | instead of get_bool_vector or get_real_vector etc | 00:37 |
@sonney2k | f-x, for which type? | 00:37 |
@sonney2k | or for all types? | 00:37 |
f-x | for all types... it is done now through a #define | 00:38 |
f-x | in StreamingSimpleFeatures | 00:38 |
f-x | get_vector could be a specialized template function for each type T | 00:38 |
f-x | and then StreamingSimpleFeatures would just use the specialized get_vector for that template type | 00:39 |
@sonney2k | f-x, wait get_vector in StreamingSimpleFeatures cannot be templated - but in the derived classes it could but then I don't think you benefit | 00:39 |
f-x | sonney2k: sorry, not that get_vector | 00:40 |
f-x | i'm confusing you | 00:40 |
f-x | i'm talking about the functions which read a vector from a file | 00:41 |
@sonney2k | ahh yes but still problem remains (despite the name) | 00:41 |
f-x | they are called get_word_vector, get_int_vector, etc.. | 00:41 |
@sonney2k | currently - yes | 00:41 |
@sonney2k | so what would you do now instead? | 00:42 |
@sonney2k | write one templated function for that | 00:42 |
f-x | call it read_vector() or something which doesn't clash | 00:42 |
f-x | yes | 00:42 |
@sonney2k | and then call it in the get_*_vector? | 00:42 |
f-x | wait | 00:42 |
f-x | it's becoming roundabout | 00:43 |
f-x | #define SET_VECTOR_READER(sg_type, sg_function)\ | 00:43 |
f-x | template <> void CStreamingSimpleFeatures<sg_type>::set_vector_reader() \ | 00:43 |
f-x | {\ | 00:43 |
f-x | parser.set_read_vector(&CStreamingFile::sg_function);\ | 00:43 |
f-x | } | 00:43 |
f-x | 00:43 | |
f-x | SET_VECTOR_READER(bool, get_bool_vector); | 00:43 |
@sonney2k | I mean I would wish one could have a templated get_vector function in the base class - but in the derived classes you still would need to overload each templated variant | 00:43 |
f-x | SET_VECTOR_READER(char, get_char_vector); | 00:43 |
f-x | SET_VECTOR_READER(int8_t, get_int8_vector); | 00:43 |
f-x | 00:43 | |
f-x | that's part of the code which sets the read functions in StreamingSimpleFeatures | 00:43 |
f-x | sonney2k: yes, you're right | 00:44 |
f-x | some implementation would have to be given for each T | 00:44 |
@sonney2k | yeah so I think you cannot do much - I think the most clean thing you could do is to use a templated function in the derived class | 00:44 |
@sonney2k | and have that called from the get_*_vector() functions | 00:44 |
@sonney2k | then at least the function code is not hidden in the macro | 00:45 |
blackburn | base/class_list.cpp: In function ‘shogun::CSGObject* __new_CLandmarkEIsomap(shogun::EPrimitiveType)’: | 00:45 |
blackburn | base/class_list.cpp:287: error: expected type-specifier before ‘CLandmarkEIsomap’ | 00:45 |
blackburn | sonney2k: any ideas? | 00:45 |
@sonney2k | blackburn, did you do the usual distclean? | 00:45 |
blackburn | will try again but seems it won't help | 00:46 |
@sonney2k | f-x, btw - the interface for streamingfeatures is much better now | 00:46 |
* sonney2k becomes *very* sleepy | 00:46 | |
f-x | sonney2k: thanks.. i've begun to like it too :) | 00:47 |
f-x | i'll try to integrate the labels into the SimpleStreamFromSimpleFeatures-like classes soon | 00:47 |
@sonney2k | f-x, so what do you think about the splitting up of StreamingFile? | 00:48 |
f-x | sonney2k: your idea is sensible, i agree | 00:48 |
blackburn | sonney2k: haven't helped | 00:48 |
f-x | there wasn't much to gain by using the current scheme anyway | 00:49 |
@sonney2k | f-x, yeah... | 00:49 |
f-x | but our names are becoming more java-ish day by day | 00:49 |
@sonney2k | blackburn, hmmhh I suspect you forgot an include somewhere | 00:50 |
@sonney2k | f-x, how so? | 00:50 |
f-x | CStreamingFileFromSimpleFeatures | 00:50 |
f-x | that looks like a java trademark | 00:50 |
@sonney2k | f-x, well I don't have an idea how to shorten this | 00:50 |
f-x | sonney2k: me neither.. just saying | 00:51 |
@sonney2k | it is only meaningful this way | 00:51 |
@sonney2k | ok | 00:51 |
blackburn | sonney2k: CEIsomap is defined in EIsomap.h and CLandmarkEIsomap here too | 00:52 |
@sonney2k | f-x, so please do this split up then - I think the code will be even more readable then :) | 00:53 |
blackburn | I have similar strange problem while working on ContingencyTableEvaluation | 00:53 |
bettyboo | he | 00:53 |
f-x | sonney2k: sure.. expect it real soon | 00:53 |
@sonney2k | blackburn, yeah but you forgot an include there too | 00:53 |
@sonney2k | (at least that is what I remember) | 00:53 |
@sonney2k | blackburn, btw what is EIsomap? | 00:54 |
@sonney2k | f-x thanks! | 00:54 |
f-x | sonney2k: btw will we be using CLabels for StreamingFeatures too? | 00:54 |
blackburn | sonney2k: E-Isomap | 00:54 |
blackburn | well there are two isomap 'ways' | 00:54 |
blackburn | to connect edges with weights less than epsilon - E-Isomap | 00:55 |
blackburn | or to connect just k neighors - K-Isomap | 00:55 |
@sonney2k | f-x, well return just a float64_t - for simplestreamingfeatures you should expect a simplefeature object and labels (optional) | 00:55 |
@sonney2k | blackburn, I see | 00:55 |
f-x | sonney2k: ok.. so i won't encapsulate everything in a CLabels object yet | 00:56 |
@sonney2k | f-x, it doesn't really make sense - you only ever return a scalar (at least now) | 00:58 |
@sonney2k | f-x, and btw. for StreamingAsciiFile you could specify the CAsciFile in constructor and use CAsciiFile's functions | 00:59 |
blackburn | I should think before coding | 01:00 |
f-x | sonney2k: for reading the file? but CAsciiFile reads the complete file, right? | 01:00 |
@sonney2k | f-x, you can use its read_line function | 01:01 |
@sonney2k | (or whatever it is called) | 01:01 |
f-x | sonney2k: it's being used in StreamingAsciiFile through CAsciiFile::getline right now, though it could be done in a better way | 01:03 |
f-x | i'll modify it a bit | 01:03 |
@sonney2k | yeah meant that function | 01:03 |
@sonney2k | ok I have to sleep now | 01:03 |
@sonney2k | cu all! | 01:04 |
f-x | see you! | 01:04 |
blackburn | see ya | 01:06 |
-!- f-x [~user@117.192.197.13] has quit [Remote host closed the connection] | 01:41 | |
CIA-32 | shogun: Sergey Lisitsyn master * r398048d / (11 files in 2 dirs): Various fixes and changes for Isomap - http://bit.ly/mKx7Dp | 01:52 |
CIA-32 | shogun: Sergey Lisitsyn master * r427a6c6 / (11 files in 2 dirs): Various fixes and changes for Isomap - http://bit.ly/kKNzAV | 01:54 |
CIA-32 | shogun: Sergey Lisitsyn master * rb8eca36 / : | 01:57 |
CIA-32 | shogun: Merge branch 'master' of ssh://shogun-toolbox.org/shogun | 01:57 |
CIA-32 | shogun: Conflicts: | 01:57 |
CIA-32 | shogun: src/modular/Preprocessor.i - http://bit.ly/jj54fw | 01:57 |
blackburn | ah sorry I forgot to amend commit before push | 01:59 |
-!- blackburn [~blackburn@31.28.40.202] has quit [Quit: Leaving.] | 02:00 | |
-!- alesis-novik [~alesis@188.74.87.206] has quit [Ping timeout: 252 seconds] | 02:33 | |
-!- alesis-novik [~alesis@vpn2-244.vpn.net.ed.ac.uk] has joined #shogun | 02:45 | |
-!- in3xes [~in3xes@210.212.58.111] has quit [Ping timeout: 240 seconds] | 06:28 | |
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 06:41 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 06:47 | |
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 06:57 | |
-!- heiko [~heiko@main.uni-duisburg.de] has joined #shogun | 10:01 | |
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 10:19 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 10:22 | |
-!- in3xes1 [~in3xes@180.149.49.227] has joined #shogun | 10:38 | |
-!- in3xes_ [~in3xes@180.149.49.227] has quit [Ping timeout: 244 seconds] | 10:42 | |
heiko | sonney2k are you there? | 10:50 |
heiko | i think I found a bug in DynArray: the operator = has a problem if the array that is to be copied is much larger than the array that is copied into. | 10:51 |
heiko | will try to fix it. | 10:51 |
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 10:54 | |
-!- in3xes1 [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 10:57 | |
-!- in3xes_ [~in3xes@180.149.49.227] has quit [Ping timeout: 252 seconds] | 11:02 | |
-!- blackburn [~blackburn@31.28.40.202] has joined #shogun | 11:22 | |
CIA-32 | shogun: Soeren Sonnenburg master * r827f2b0 / (21 files in 6 dirs): | 11:36 |
CIA-32 | shogun: Merge pull request #153 from karlnapf/master | 11:36 |
CIA-32 | shogun: new subset implementation (+7 more commits...) - http://bit.ly/jqI4sV | 11:36 |
heiko | sonney2k are you there? | 11:58 |
blackburn | heiko: I can merge your pull | 12:26 |
blackburn | *if you want* :) | 12:27 |
heiko | hej blackburn :) | 12:37 |
heiko | yes, if you think its ok | 12:37 |
heiko | I am not sure whether soeren likes the SG_CALLOC stuff. The bugfix should be ok | 12:39 |
heiko | will go for lunch now | 12:39 |
heiko | see you :) | 12:39 |
blackburn | heiko: ok 'll wait for Soeren :) | 12:43 |
blackburn | enjoy your meal | 12:43 |
-!- heiko [~heiko@main.uni-duisburg.de] has quit [Ping timeout: 258 seconds] | 12:51 | |
blackburn | hmm I guess some changes caused really slow compilation | 13:18 |
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 13:27 | |
-!- in3xes_ [~in3xes@210.212.58.111] has joined #shogun | 13:36 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 276 seconds] | 13:39 | |
-!- alesis-novik [~alesis@vpn2-244.vpn.net.ed.ac.uk] has quit [Ping timeout: 246 seconds] | 13:44 | |
-!- heiko [~heiko@main.uni-duisburg.de] has joined #shogun | 13:52 | |
-!- blackburn [~blackburn@31.28.40.202] has quit [Quit: Leaving.] | 13:56 | |
-!- alesis-novik [~alesis@188.74.87.206] has joined #shogun | 13:59 | |
-!- sploving1 [~sploving@2001:cc0:2020:2011:5eff:35ff:fe04:f091] has joined #shogun | 14:24 | |
sploving1 | anyone here? | 14:30 |
sploving1 | ./preprocessor/ClassicIsomap.h:50: error: ‘m_target_dim’ was not declared in this scope | 14:31 |
sploving1 | base/class_list.cpp: In function ‘shogun::CSGObject* __new_CLandmarkIsomap(shogun::EPrimitiveType)’: | 14:31 |
sploving1 | base/class_list.cpp:436: error: cannot convert ‘shogun::CLandmarkIsomap*’ to ‘shogun::CSGObject*’ in return | 14:31 |
sploving1 | base/class_list.cpp: In function ‘shogun::CSGObject* __new_CClassicIsomap(shogun::EPrimitiveType)’: | 14:31 |
sploving1 | base/class_list.cpp:439: error: cannot convert ‘shogun::CClassicIsomap*’ to ‘shogun::CSGObject*’ in return | 14:31 |
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 14:34 | |
sploving1 | I merged upstream and configure java_modular | 14:34 |
sploving1 | it shows the above error | 14:35 |
sploving1 | any one could help? | 14:35 |
heiko | hi sploving | 14:39 |
heiko | sorry, I do not have java configured | 14:39 |
heiko | currently struggeling with java | 14:39 |
heiko | ehm python_modula | 14:39 |
heiko | r | 14:39 |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 260 seconds] | 14:41 | |
sploving1 | I think maybe someone changes for Isomap | 14:42 |
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 14:49 | |
-!- in3xes__ [~in3xes@180.149.49.227] has joined #shogun | 14:53 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 246 seconds] | 14:57 | |
-!- in3xes [~in3xes@180.149.49.227] has joined #shogun | 15:00 | |
-!- in3xes__ [~in3xes@180.149.49.227] has quit [Ping timeout: 252 seconds] | 15:03 | |
-!- in3xes__ [~in3xes@180.149.49.227] has joined #shogun | 15:04 | |
-!- sploving1 [~sploving@2001:cc0:2020:2011:5eff:35ff:fe04:f091] has left #shogun [] | 15:05 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 258 seconds] | 15:07 | |
-!- in3xes__ [~in3xes@180.149.49.227] has quit [Ping timeout: 255 seconds] | 15:12 | |
-!- in3xes__ [~in3xes@180.149.49.227] has joined #shogun | 15:25 | |
-!- heiko [~heiko@main.uni-duisburg.de] has quit [Ping timeout: 258 seconds] | 15:28 | |
-!- in3xes__ is now known as in3es | 15:32 | |
-!- in3es is now known as in3xes | 15:32 | |
-!- in3xes__ [~in3xes@180.149.49.227] has joined #shogun | 15:34 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 244 seconds] | 15:37 | |
-!- in3xes1 [~in3xes@180.149.49.227] has joined #shogun | 15:37 | |
-!- in3xes__ [~in3xes@180.149.49.227] has quit [Ping timeout: 252 seconds] | 15:39 | |
-!- in3xes_ [~in3xes@210.212.58.111] has quit [Ping timeout: 276 seconds] | 15:41 | |
-!- f-x [~user@117.192.219.76] has joined #shogun | 16:13 | |
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 16:15 | |
-!- in3xes1 [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 16:19 | |
-!- in3xes_ is now known as in3xes | 16:25 | |
-!- heiko [~heiko@main.uni-duisburg.de] has joined #shogun | 17:11 | |
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 17:20 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 263 seconds] | 17:24 | |
-!- in3xes_ is now known as in3xes | 17:30 | |
-!- f-x [~user@117.192.219.76] has quit [Read error: Connection reset by peer] | 17:39 | |
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 17:39 | |
-!- heiko [~heiko@main.uni-duisburg.de] has quit [Ping timeout: 258 seconds] | 17:42 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 17:43 | |
-!- in3xes_ is now known as in3xes | 17:43 | |
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 17:50 | |
-!- in3xes [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 17:54 | |
-!- in3xes1 [~in3xes@59.163.196.121] has joined #shogun | 17:56 | |
-!- in3xes_ [~in3xes@180.149.49.227] has quit [Ping timeout: 240 seconds] | 17:59 | |
-!- f-x [~user@117.192.219.76] has joined #shogun | 18:06 | |
-!- in3xes1 is now known as in3xes | 18:43 | |
-!- blackburn [~blackburn@31.28.40.202] has joined #shogun | 19:26 | |
blackburn | oh I forgot some includes.. how it was working on my | 19:33 |
blackburn | sonney2k: some problem with heiko's patch: it is not compiling now :) | 20:35 |
CIA-32 | shogun: Sergey Lisitsyn master * rc47b8ec / src/libshogun/preprocessor/Isomap.h : Fixed includes at Isomap.h - http://bit.ly/iYQux2 | 20:48 |
blackburn | eh... everything is ok. | 21:05 |
blackburn | wtf | 21:05 |
-!- blackburn [~blackburn@31.28.40.202] has quit [Ping timeout: 255 seconds] | 21:12 | |
@sonney2k | all good now? | 21:18 |
CIA-32 | shogun: Soeren Sonnenburg master * re5b3696 / src/java_modular/swig_typemaps.i : | 21:27 |
CIA-32 | shogun: Merge pull request #155 from sploving/master | 21:27 |
CIA-32 | shogun: support SGVector for UJMP - http://bit.ly/mLWtUq | 21:27 |
-!- blackburn [~blackburn@31.28.40.202] has joined #shogun | 21:28 | |
blackburn | sonney2k: around? | 21:30 |
@sonney2k | yes | 21:30 |
@sonney2k | all good now? | 21:31 |
blackburn | sonney2k: nope | 21:31 |
blackburn | sonney2k: is it compiling on your machine? | 21:31 |
@sonney2k | ? | 21:31 |
@sonney2k | cannot test right now | 21:31 |
blackburn | really strange errors | 21:31 |
blackburn | ../shogun/features/Features.h:282: error: ‘shogun::CSubset* shogun::CFeatures::m_subset’ is protected | 21:31 |
blackburn | ../shogun/features/StringFeatures.h:1685: error: within this context | 21:31 |
blackburn | ../shogun/features/Features.h: In member function ‘bool shogun::CStringFeatures<ST>::obtain_from_char_features(shogun::CStringFeatures<CT>*, int32_t, int32_t, int32_t, bool) [with CT = char, ST = int]’: | 21:31 |
@sonney2k | what does not work? | 21:31 |
blackburn | the most strange thing about it - it fails on wrappers compilation | 21:32 |
@sonney2k | hmmh i cannot make sense of this | 21:33 |
blackburn | sonney2k: I guess I know | 21:34 |
blackburn | may be subset.h should be imported? | 21:34 |
CIA-32 | shogun: Soeren Sonnenburg master * rb803c05 / (8 files in 4 dirs): | 21:36 |
CIA-32 | shogun: Merge pull request #154 from karlnapf/master | 21:36 |
CIA-32 | shogun: bug in DynArray - http://bit.ly/l7sCtm | 21:36 |
@sonney2k | blackburn, in Features_includes.i | 21:37 |
@sonney2k | and Features.i | 21:37 |
@sonney2k | makes sense | 21:37 |
@sonney2k | and rename too | 21:37 |
blackburn | sonney2k: I have added to stirng features, compiling now | 21:37 |
blackburn | will check in a min | 21:37 |
@sonney2k | f-x, any news - are you working on the patch | 21:39 |
@sonney2k | ? | 21:39 |
blackburn | nope | 21:39 |
f-x | sonney2k: yeah... actually had some clarifications to make | 21:40 |
f-x | what functions actually go into the SimpleStream interface? | 21:40 |
@sonney2k | the vector ones | 21:41 |
f-x | get_int_vector(), get_real_vector() etc., right? | 21:42 |
f-x | and get_*_vector_and_label()...? | 21:42 |
blackburn | sonney2k: import haven't helped. | 21:44 |
@sonney2k | blackburn, no import! | 21:45 |
@sonney2k | include | 21:45 |
blackburn | include yes | 21:45 |
@sonney2k | baby crying... | 21:45 |
@sonney2k | l8r | 21:45 |
blackburn | now should be fixed | 21:57 |
blackburn | heiko did it indeed wrong | 21:58 |
blackburn | it should not work | 21:58 |
blackburn | there was an access to protected field | 21:58 |
CIA-32 | shogun: Sergey Lisitsyn master * rcc04bff / src/libshogun/features/StringFeatures.h : Added FIXME and fixed StringFeatures compile bug - http://bit.ly/mkHgJE | 22:47 |
* blackburn is translating FAQ to russian hehe | 22:57 | |
@sonney2k | blackburn, I think heiko's code is right | 23:17 |
blackburn | sonney2k: it is not compilable | 23:17 |
@sonney2k | m_subset is protected yes but it is defined in CFeatures and stringfeatures derives from that | 23:17 |
@sonney2k | so it can access m_subset | 23:18 |
blackburn | sonney2k: it is passed as parameter | 23:18 |
blackburn | it is ok for g++ but swig fails on that | 23:18 |
blackburn | sonney2k: not just m_subset but m_subset of StringFeatures passed to that method as parameter I mean | 23:19 |
@sonney2k | hmmhh I thought it doesn' t matter if this is a parameter | 23:19 |
blackburn | sf->m_subset | 23:19 |
@sonney2k | yes but I don't see what could be wrong here | 23:20 |
blackburn | just try to compile it | 23:20 |
blackburn | I even re-cloned shogun :D | 23:20 |
blackburn | test | 23:31 |
@sonney2k | blackburn, I think this is because of the template thing | 23:32 |
@sonney2k | I mean otherwise lines such as sf->m_subset->subset_idx_conversion(i); would also cause errors (sf is CStringFeatures<ST>* sf) | 23:33 |
blackburn | sonney2k: don't know, the only thing I wonder - is it compiles on heiko's machine? :) | 23:33 |
bettyboo | not bad!?! | 23:33 |
@sonney2k | yeah that is strange | 23:34 |
blackburn | fuck.. my K-Isomap is wrong. again! | 23:34 |
blackburn | sonney2k: what would you do if you was needed to remove NOT k-nearest neighbor edges from distance matrix? | 23:37 |
blackburn | -> infty them | 23:37 |
@sonney2k | makes sense | 23:38 |
blackburn | I mean my very naive way to do that | 23:38 |
blackburn | sort row(or column) indexes | 23:38 |
blackburn | with decreasing distance | 23:38 |
blackburn | and just -> infty starting from k+1 | 23:39 |
blackburn | the most surprising thing about that - this will produce NONsymmetric matrix :( | 23:40 |
blackburn | hmm may be just 'symmetrize' it.. | 23:44 |
blackburn | I really wonder I'm stucked on such a EASY thing :D | 23:49 |
--- Log closed Thu Jun 30 00:00:31 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!