Zeitpunkt Nutzer Delta Tröts TNR Titel Version maxTL Sa 17.08.2024 00:00:04 2.051.989 +1.268 100.767.258 49,1 Mastodon 4.3.0... 500 Fr 16.08.2024 00:00:00 2.050.721 +1.111 100.660.691 49,1 Mastodon 4.3.0... 500 Do 15.08.2024 00:00:01 2.049.610 +1.147 100.539.771 49,1 Mastodon 4.3.0... 500 Mi 14.08.2024 00:00:05 2.048.463 0 100.426.123 49,0 Mastodon 4.3.0... 500 Mi 14.08.2024 00:00:05 2.048.463 +1.555 100.426.123 49,0 Mastodon 4.3.0... 500 Di 13.08.2024 00:00:00 2.046.908 +1.369 100.311.382 49,0 Mastodon 4.3.0... 500 Mo 12.08.2024 00:00:00 2.045.539 +1.270 100.192.858 49,0 Mastodon 4.3.0... 500 So 11.08.2024 00:00:10 2.044.269 +1.003 100.082.452 49,0 Mastodon 4.3.0... 500 Sa 10.08.2024 00:00:04 2.043.266 +1.405 99.975.166 48,9 Mastodon 4.3.0... 500 Fr 09.08.2024 00:00:03 2.041.861 0 99.897.120 48,9 Mastodon 4.3.0... 500
Joxean Koret (@matalaz) (@joxean) · 04/2017 · Tröts: 15.728 · Folger: 1.902
Sa 17.08.2024 11:49
With this data, I think I know which algorithm is the best fit for my use-case.
BTW, I'm surprised with the difference in times between the top 4 performing algorithms: most algorithms take more than 1 hour, but the decision trees classifier takes around 12 minutes.
LibreOffice spreadsheet showing the following tabular data: Algorithm Name Accuracy on testing data Accuracy on training dataset Deviation Duration ExtraTreesClassifier 0,97 0,91 0,01 ~01:15:00 RandomForestClassifier 0,97 0,92 0,01 +2 hours BaggingClassifier 0,96 0,91 0,01 ~01:10:00 DecisionTreeClassifier 0,95 0,88 0,02 12 minutes KNeighborsClassifier 0,94 0,88 0,01 ~2:30:00 MLPClassifier 0,91 0,9 0,01 ~5 hours GradientBoostingClassifier 0,9 0,91 0,01 ~3 hours AdaBoostClassifier 0,89 0,9 0,02 ~40 minutes SGDClassifier 0,85 0,85 0,05 10 minutes LinearDiscriminantAnalysis 0,85 0,88 0,02 2 minutes (!) LogisticRegression 0,85 0,88 0,01 3 minutes BernoulliNB 0,83 0,8 0,02 1 minute GaussianNB 0,82 0,86 0,02 ~2 minutes BayesianRidge 0,31 0,43 0,06 1 minute
[Öffentlich] Antw.: 0 Wtrl.: 0 Fav.: 0