--- Log opened Mon Sep 14 00:00:19 2015 | ||
-!- shaochuan [~shaochuan@2601:647:4603:2a70:18fb:2b68:9e:929b] has joined #shogun | 02:03 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:18fb:2b68:9e:929b] has quit [Remote host closed the connection] | 02:07 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:4c46:2aa9:3797:7755] has joined #shogun | 02:14 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:4c46:2aa9:3797:7755] has quit [Remote host closed the connection] | 02:17 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 02:17 | |
-!- uchihaitachi08_ [~sameer@203.110.243.23] has quit [Remote host closed the connection] | 03:52 | |
-!- uchihaitachi08_ [~sameer@203.110.243.23] has joined #shogun | 03:53 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Remote host closed the connection] | 04:34 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 04:35 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Remote host closed the connection] | 04:38 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:4c46:2aa9:3797:7755] has joined #shogun | 04:39 | |
-!- shaochua_ [~shaochuan@2601:647:4603:2a70:d467:a2c3:156e:5d67] has joined #shogun | 04:40 | |
-!- shaochua_ [~shaochuan@2601:647:4603:2a70:d467:a2c3:156e:5d67] has quit [Remote host closed the connection] | 04:41 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:4c46:2aa9:3797:7755] has quit [Ping timeout: 244 seconds] | 04:43 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:949b:eb15:dfc0:34d] has joined #shogun | 04:44 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:949b:eb15:dfc0:34d] has quit [Remote host closed the connection] | 04:45 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 04:47 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Remote host closed the connection] | 04:48 | |
shogun-buildbot | build #1050 of nightly_default is complete: Failure [failed test notebooks] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/1050 blamelist: Bj?rn Esser <bjoern.esser@gmail.com>, Robert Pollak <robert.pollak@mykolab.com> | 05:58 |
---|---|---|
-!- uchihaitachi08_ [~sameer@203.110.243.23] has quit [Remote host closed the connection] | 06:14 | |
-!- uchihaitachi08_ [~sameer@203.110.243.23] has joined #shogun | 06:14 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 06:36 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Remote host closed the connection] | 06:55 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 07:16 | |
-!- uchihaitachi08_ [~sameer@203.110.243.23] has quit [Remote host closed the connection] | 08:49 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Remote host closed the connection] | 08:54 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:9046:7879:f945:dc53] has joined #shogun | 09:00 | |
-!- shaochuan [~shaochuan@2601:647:4603:2a70:9046:7879:f945:dc53] has quit [Remote host closed the connection] | 09:04 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has joined #shogun | 09:05 | |
-!- shaochuan [~shaochuan@c-50-184-81-180.hsd1.ca.comcast.net] has quit [Ping timeout: 246 seconds] | 09:10 | |
-!- ddk_ [~danny@2a02:1812:518:7300:dc40:f7bc:57dc:27c5] has joined #shogun | 10:55 | |
ddk_ | I succeeded to compile shogun V4 under ubuntu 14.04.3LTS | 10:56 |
ddk_ | but some libs are missing like eigen3 any idea where to find those? | 10:57 |
ddk_ | it's ok it is called libeigen3-dev, I am recompiling now :-) | 11:01 |
-!- ddk_ [~danny@2a02:1812:518:7300:dc40:f7bc:57dc:27c5] has quit [Ping timeout: 256 seconds] | 11:08 | |
-!- ddk_ [~danny@2a02:1812:518:7300:dc40:f7bc:57dc:27c5] has joined #shogun | 11:20 | |
ddk_ | mmm still no output on the example regression_gaussian_process_gaussian | 11:30 |
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 256 seconds] | 13:26 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 13:28 | |
-!- mode/#shogun [+o besser82] by ChanServ | 13:28 | |
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 264 seconds] | 13:33 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 13:39 | |
-!- mode/#shogun [+o besser82] by ChanServ | 13:39 | |
ddk_ | Any idea why even with the libeigen3-dev the regression_gaussian_process_gaussian is not running??? | 15:21 |
ddk_ | any idea what following 0 means in #if defined(HAVE_EIGEN3) && defined(HAVE_NLOPT) && 0 | 15:41 |
@wiking | dkk | 15:42 |
@wiking | dkk_ | 15:42 |
@wiking | it's because // temporally disabled, since API was changed | 15:42 |
ddk_ | yes wiking | 15:43 |
ddk_ | so the samples don't work? | 15:43 |
@wiking | yeah api has changed since this has been written | 15:44 |
ddk_ | a pity because I want to create the same test as on your website | 15:47 |
ddk_ | gaussian process regression with ml2 learning | 15:47 |
ddk_ | because I have a list of stock closing prices I want to use as test data to predict the next days in the future | 15:48 |
lisitsyn | probably you can use older version? | 15:49 |
lisitsyn | I am out of this change, wiking who's doing that? | 15:49 |
lisitsyn | wu? | 15:49 |
ddk_ | thanks lisitsyn any idea what would be the best model to do this? | 15:50 |
ddk_ | I have tried NN but the results are not good | 15:51 |
lisitsyn | ddk_: what about support vector regression? | 15:51 |
ddk_ | seems also ok but I can't find any example how to use it, I am strong in C/C++ but machine learning is all new to me ... | 15:53 |
ddk_ | lisitsyn do you have a example of this in c/c++? | 15:55 |
lisitsyn | ddk_: yeah check e.g. regression_svrlight | 15:57 |
ddk_ | lisitsyn: can't find it in the examples only libsvr | 15:58 |
@wiking | ddk_: examples/undocumented/libshogun/regression_libsvr.cpp | 15:59 |
ddk_ | wiking: when I run this one I get mean squared error: 0.033732 nothing more | 16:00 |
lisitsyn | yes, what else? ;) | 16:02 |
ddk_ | lisitsyn: sorry you guys are experts in this I am feeling a bit a baby ;-) | 16:02 |
lisitsyn | ddk_: ha not a problem | 16:03 |
lisitsyn | ddk_: its a simple example | 16:03 |
lisitsyn | just runs and evaluates what it gets | 16:03 |
lisitsyn | we expect user to read it and learn by example | 16:03 |
ddk_ | I just want to feed the model 100 points to learn and then to predict the next few points | 16:04 |
ddk_ | like the cool example on the website | 16:04 |
ddk_ | lisitsyn: I think I will look for a easier toolbox because this one has so many features I can't grasp it ... | 16:08 |
lisitsyn | ddk_: I guess its sklearn | 16:09 |
lisitsyn | :) | 16:09 |
ddk_ | sklearn? | 16:09 |
lisitsyn | yes that's the usual choice nowadays | 16:09 |
lisitsyn | its python though | 16:09 |
ddk_ | lisitsyn: the problem is I need only 1 solution but this toolbox is so big I have no clue where to begin. The website is very cool with the try it yourself pages but those examples you can't find back in the toolbox ... | 16:11 |
ddk_ | lisitsyn: I just need to replace the toydata with my own data points and I am ready to go :-) | 16:12 |
ddk_ | lisitsyn: scikit seems nice to but indeed python and the results are no so good as Shogun with the same data points | 16:15 |
ddk_ | not a lot of svm toolboxes support gaussian kernel | 16:33 |
lisitsyn | ddk_: sorry, back | 16:52 |
lisitsyn | I don't think result would be very different | 16:52 |
ddk_ | I just tested it and Shogun give a better result ... | 16:53 |
lisitsyn | ddk_: why do you want to use C++? :) | 16:53 |
lisitsyn | ah | 16:53 |
lisitsyn | ddk_: so you compare sklearn and shogun on same data? | 16:53 |
ddk_ | because I am a long time c/c++ user | 16:53 |
ddk_ | yep | 16:53 |
ddk_ | real stock market data of IBM :-) | 16:53 |
lisitsyn | heh that's cool | 16:54 |
lisitsyn | what method it is? | 16:54 |
ddk_ | svr with gaussian but I just read a paper and radial basis seems to give better results | 17:00 |
lisitsyn | ok cool | 17:01 |
ddk_ | I found LIBSVM, compiles very fast | 17:01 |
lisitsyn | yes we have libsvm | 17:01 |
ddk_ | and what are the results? | 17:01 |
lisitsyn | well should be the same | 17:02 |
lisitsyn | its just the same libsvm put inside the library | 17:02 |
ddk_ | do you have any sample data files for libsvm? | 17:02 |
lisitsyn | ddk_: you mean in libsvm format? | 17:03 |
ddk_ | yep | 17:03 |
lisitsyn | uhmm searching | 17:05 |
lisitsyn | ddk_: on the libsvm website | 17:05 |
lisitsyn | there are a few datasets | 17:05 |
lisitsyn | you can read them with libsvm itself and shogun | 17:06 |
lisitsyn | if you want to compare | 17:06 |
ddk_ | found it, but a simple data file is hard to find :-) | 17:09 |
-!- besser82 [~besser82@fedora/besser82] has quit [Ping timeout: 256 seconds] | 17:13 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 17:13 | |
-!- mode/#shogun [+o besser82] by ChanServ | 17:13 | |
ddk_ | lisitsyn: RBF kernel ... | 17:14 |
lisitsyn | ddk_: what? :) | 17:15 |
ddk_ | best results | 17:15 |
lisitsyn | ddk_: well gaussian aka rbf is good | 17:15 |
ddk_ | and scaling the data makes the results even better | 17:16 |
lisitsyn | ddk_: expected | 17:17 |
-!- ddk_ [~danny@2a02:1812:518:7300:dc40:f7bc:57dc:27c5] has quit [Quit: Leaving] | 18:48 | |
--- Log closed Tue Sep 15 00:00:20 2015 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!