Commit Graph

55 Commits

Author SHA1 Message Date
Jonas Weinz
57bb25e558 more evaluation 2018-07-23 13:25:37 +02:00
Jonas Weinz
953ca76e6b evaluation on sentiment dataset 2018-07-23 09:23:17 +02:00
Jonas Weinz
8823582d4b simple improvements 2018-07-22 10:55:31 +02:00
Jonas Weinz
96e4606880 added ability to filter by sentence length 2018-07-21 10:32:34 +02:00
Jonas Weinz
1c8c15b0d5 plotting validation error 2018-07-21 09:48:25 +02:00
Jonas Weinz
3f5b5fe3f3 sync cont. learner 2018-07-20 13:41:13 +02:00
Jonas Weinz
b4eb220973 added possibility to filter by emoji list 2018-07-20 12:49:55 +02:00
Jonas Weinz
092c18ba16 continous lerner 2018-07-19 17:47:27 +02:00
Jonas Weinz
f4766912e1 dump put emoji counts 2018-07-19 16:46:19 +02:00
Jonas Weinz
8300ed4510 filtered out empty labels 2018-07-19 15:58:06 +02:00
Jonas Weinz
16551c64f6 simple fix 2018-07-10 16:23:54 +02:00
Jonas Weinz
7f6c9791ae simple improvements 2018-07-10 15:59:28 +02:00
Jonas Weinz
b4ae0b033e added possibility to load pretrained doc2vecD 2018-07-10 13:48:51 +02:00
aikira7
951b552f61 doc2vec Model trainiert auf Brown Corpus 2018-07-09 18:28:29 +02:00
Jonas Weinz
abd61c35d1 hotfix 2018-07-03 15:41:39 +02:00
Jonas Weinz
cffa49eade hotfix 2018-07-03 15:40:04 +02:00
Jonas Weinz
8cbd7cbe90 ability to set batch sizes to zero 2018-07-03 15:37:15 +02:00
Jonas Weinz
84aa7f288f some keras batch size fixes 2018-07-03 15:07:45 +02:00
Jonas Weinz
8abebc5957 integrated word to vec 2018-07-03 14:20:31 +02:00
Jonas Weinz
55bd7a9b3a minor fixes 2018-06-26 20:47:33 +02:00
Jonas Weinz
480c1e8928 fixed missing return 2018-06-26 20:33:55 +02:00
Jonas Weinz
b8bd92db06 lemmatization and stemming in prediction 2018-06-26 20:07:37 +02:00
Jonas Weinz
2a6a29b88b bugfixes and improvements in sample handling, working on sentiment mean as labels 2018-06-26 16:27:27 +02:00
Jonas Weinz
10fd5817e7 updated notebook according to changes in the python file 2018-06-26 14:47:39 +02:00
Jonas Weinz
df811196e5 continous learning with gui works for keras models 2018-06-26 14:44:05 +02:00
Jonas Weinz
3dcc18221f batch fitting with sparse matrices and keras is very uncomfortable *sigh* 2018-06-25 19:17:38 +02:00
Jonas Weinz
7e3b5be4fe removed interactive parts from simple_twitter_learning python module 2018-06-25 11:03:12 +02:00
Jonas Weinz
b067c789a7 merged kmeans approach 2018-06-20 18:10:54 +02:00
Jonas Weinz
0bedb6060d documentation and python file for simple_twitter_learning 2018-06-20 13:38:18 +02:00
Jonas Weinz
a623322ebb just very little tweaks 2018-06-20 11:34:36 +02:00
Jonas Weinz
8425355783 refactoring simple twitter learning done 2018-06-19 16:29:11 +02:00
Jonas Weinz
615c0ef330 Merge branch 'master' into refactoring 2018-06-19 16:25:19 +02:00
Jonas Weinz
e0ed251c8b first refactored twitter_learning version 2018-06-19 16:23:28 +02:00
Jonas Weinz
dd95cfa144 just some little updates 2018-06-16 18:04:49 +02:00
Carsten
32d66c2d0a more changes for output 2018-06-16 17:55:27 +02:00
Carsten
07ce457056 first evaluation notebook 2018-06-16 15:36:14 +02:00
Jonas Weinz
e45a7f5531 fixed table 2018-06-16 15:34:30 +02:00
Jonas Weinz
a45fa5f843 printing out sentiment vector in the testing area 2018-06-12 16:41:09 +02:00
aikira7
3a50765954 added preprocessing option 2018-06-12 10:32:56 +02:00
Jonas Weinz
6355736b99 added direct download links in notebook 2018-06-11 21:49:56 +02:00
Jonas Weinz
0a97b9aff1 added some statistical evaluation 2018-06-11 21:26:49 +02:00
Jonas Weinz
63a838df5e first not bad looking keras result (forgot to update notebook) 2018-06-11 20:14:21 +02:00
Jonas Weinz
d9f216346e first not bad looking keras result 2018-06-11 20:13:27 +02:00
Jonas Weinz
598d0f8315 it's getting a little bit better... 2018-06-11 18:19:52 +02:00
Jonas Weinz
e2d793297f only using top emojis 2018-06-11 18:04:39 +02:00
Jonas Weinz
35a963d608 hotfix fo wrong regex pattern 2018-06-11 17:02:54 +02:00
Jonas Weinz
ee0ba46310 little fixes 2018-06-10 18:32:41 +02:00
Jonas Weinz
8fd29ab33d sentiment learning 2018-06-10 17:31:43 +02:00
Jonas Weinz
5e2af0697e minor adjustments 2018-06-10 15:06:45 +02:00
Jonas Weinz
6c835bbc29 keras approach 2018-06-10 14:41:41 +02:00