MLcomp version alpha Statistics: 1213 users, 402 datasets, 328 programs, 9938 runs (0 queued, 0 running), 0 workers

movielens100k

100K MovieLens movie ratings dataset from http://www.grouplens.org/. 100,000 ratings (1-5) from 943 users on 1682 movies. Test set contains exactly 10 ratings per user. See included README.txt for more information.

Format:	CollaborativeFiltering
User:	lester
Size:	1M
Status:	processed
Access:	open
Download:	Login required!
#train	90570
#test	9430
#rows	943
#cols	0
minY
maxY

Run a program on this dataset

Select a program CollaborativeFiltering 1-25 of 107 < > Filter:

	Name	Created	User	Rating <<	#runs
run	xterm-knnsvd_shrunk_unbias_rev-r	2y84d ago	xterm	61	6
run	MyMediaLite-slopeone	217d21h ago	zenogantner	61	6
run	mean_user_rating_ruby	2y109d ago	jake	61	6
run	mean_movie_rating_ruby	2y105d ago	jake	61	6
run	MyMediaLite-user-item-baseline	216d18h ago	zenogantner	60	16
run	jjariyas-svdj-r	2y88d ago	jjariyas	60	6
run	nanhengwu-mf_for1m2-r	2y87d ago	nanhengwu	60	7
run	MyMediaLite-matrix-factorization-k-10	217d3h ago	zenogantner	60	19
run	MyMediaLite-matrix-factorization-k-20	217d3h ago	zenogantner	60	19
run	MyMediaLite-matrix-factorization-k-40	217d3h ago	zenogantner	60	19
run	MyMediaLite-matrix-factorization-k-5	217d3h ago	zenogantner	59	19
run	simple-knn-with-tuning-and-shrunk-column	323d18h ago	yorinasub17	59	7
run	whoburg-ALS_K1-r	2y87d ago	whoburg	59	6
run	cberner-KNN-r	2y91d ago	cberner	59	6
run	MyMediaLite-biased-matrix-factorization-k-60	217d3h ago	zenogantner	59	10
run	mjanderson-knn-r	2y88d ago	mjanderson	59	6
run	MyMediaLite-biased-matrix-factorization-k-40	217d3h ago	zenogantner	58	10
run	MyMediaLite-biased-matrix-factorization-k-20	217d18h ago	zenogantner	58	11
run	MyMediaLite-user-average	217d21h ago	zenogantner	58	6
run	mean_user_rating_ruby_2	2y105d ago	jake	58	6
run	lester-shrunk_item_mean-r	2y101d ago	lester	58	15
run	lester-shrunk_item_mean-testSunday	2y91d ago	jjariyas	58	15
run	MyMediaLite-biased-matrix-factorization-k-10	217d3h ago	zenogantner	58	10
run	MyMediaLite-biased-matrix-factorization-k-5	217d3h ago	zenogantner	58	10
run	cberner-KNNImproved-r	2y89d ago	cberner	58	6

1. pagination_page => 0
2. current_sort_col => avg_percentile
3. paginate => true
4. model => Program
5. name => programs
6. limit => 25
7. reverse_sort => true
8. show_footer => true
9. pagination_maxed => false
10. columns => prg_create_runprg_nameprg_created_atprg_userprg_avg_percentileprg_num_runs
11. width => 100%
12. joins =>
13. include => uservresult
14. title => Select a program
15. filters => is_helperfalseprocess_statussuccesstask_typeCollaborativeFiltering

Existing runs on movielens100k (all) 1-25 of 117 < > Filter:

ID	Program	Dataset	Tuned hyper.	User	Updated	Status	Total time	Memory	Error >>
Run #14836	self-reflective-multi-task-GP-with-side-info-r	movielens100k	no	internal	208d14h ago	done	1h20m	469M	0.923
Run #14841	self-reflective-multi-task-GP-r	movielens100k	no	internal	208d14h ago	done	1h20m	481M	0.928
Run #14792	localMTL	movielens100k	no	internal	208d14h ago	done	1h54m	424M	0.937
Run #1214	xterm-knnsvd_shrunk_unbias-r	movielens100k	no	xterm	1y122d ago	done	9m59s	425M	0.938
Run #1265	xterm-knnsvd_shrunk_unbias_rev-r	movielens100k	no	xterm	1y122d ago	done	8m28s	425M	0.938
Run #875	cberner-KNNImproved-r	movielens100k	no	cberner	1y122d ago	done	5m26s	143M	0.942
Run #14887	MyMediaLite-item-knn-pearson	movielens100k	no	internal	208d14h ago	done	7m37s	418M	0.946
Run #14877	MyMediaLite-user-knn-pearson	movielens100k	no	internal	208d14h ago	done	13m6s	418M	0.949
Run #1110	whoburg-ALS_K3_lambda2-r	movielens100k	no	whoburg	1y122d ago	done	37m54s	425M	0.950
Run #2106	mjanderson-matrix-factorization-r	movielens100k	no	internal	1y122d ago	done	2m6s	421M	0.954
Run #14857	MyMediaLite-biased-matrix-factorization-k-60	movielens100k	no	internal	208d15h ago	done	8m31s	423M	0.954
Run #14892	MyMediaLite-item-knn-cosine	movielens100k	no	internal	208d14h ago	done	9m31s	423M	0.957
Run #1175	kenghao-knn_k50-r	movielens100k	no	kenghaochang	1y122d ago	done	10m0s	425M	0.959
Run #14882	MyMediaLite-user-knn-cosine	movielens100k	no	internal	208d14h ago	done	19m47s	418M	0.959
Run #1118	whoburg-ALS_K3-r	movielens100k	no	whoburg	1y122d ago	done	1h10m	60M	0.959
Run #1154	kenghao-matrix_em_k10-r	movielens100k	no	kenghaochang	1y122d ago	done	10m2s	424M	0.959
Run #1198	kenghao-matrix_em_k10_v2-r	movielens100k	no	kenghaochang	1y122d ago	done	4m19s	246M	0.959
Run #14742	MyMediaLite-matrix-factorization-k-5	movielens100k	no	internal	208d16h ago	done	20s	417M	0.960
Run #13966	MML-BMF	movielens100k	no	zenogantner	218d15h ago	done	46s	423M	0.961
Run #13992	MyMediaLite-slopeone	movielens100k	no	zenogantner	217d20h ago	done	2m7s	423M	0.961
Run #1116	whoburg-ALS_K5-r	movielens100k	no	whoburg	1y122d ago	done	1h15m	425M	0.962
Run #14756	MyMediaLite-matrix-factorization-k-10	movielens100k	no	internal	208d16h ago	done	23s	418M	0.965
Run #1039	whoburg-ALS_K1-r	movielens100k	no	whoburg	1y122d ago	done	4m6s	416M	0.968
Run #14897	MyMediaLite-user-item-baseline	movielens100k	no	internal	208d14h ago	done	9s	422M	0.972
Run #14872	MyMediaLite-biased-matrix-factorization-k-40	movielens100k	no	internal	208d14h ago	done	6m29s	414M	0.973

1. pagination_page => 0
2. filter_field => programs.namedatasets.nameruns.idrun_statuses.status
3. current_sort_col => error
4. paginate => true
5. model => Run
6. name => runs
7. limit => 25
8. show_footer => true
9. reverse_sort => false
10. pagination_maxed => false
11. columns => run_idrun_core_programrun_core_datasetrun_hyperrun_userrun_updated_atrun_statusrun_timerun_memoryrun_error
12. width => 100%
13. joins => statuscore_programcore_dataset
14. filters => core_dataset_id341programs.is_helperfalse
15. title => Existing runs on movielens100k

Processing details

• exitCode: 0
• success: true
• test:
  • maxLabelValue: 5.0
  • minLabelValue: 1.0
  • numCols: 1664
  • numExamples: 9430
  • numRows: 943
  • success: true
  • time: 0
• train:
  • maxLabelValue: 5.0
  • minLabelValue: 1.0
  • numCols: 1682
  • numExamples: 90570
  • numRows: 943
  • success: true
  • time: 1

Comments:

• 

  Jacob Abernethy (jake)
  UC Berkeley Comp Sci
  Posted at 11/15/2009

  Netflix...

  Oriol - you should have joined Lester's netflix team. His team basically tied for first place, but didn't get the million because their submission was slightly later.
• 

  Oriol Vinyals (xterm)
  UC Berkeley
  Posted at 11/14/2009

  :)

  I did KNN plus SVD. I wanted to do FA but I'd rather use C++ to do that... R is too slow!
• 

  Jacob Abernethy (jake)
  UC Berkeley Comp Sci
  Posted at 11/14/2009

  Wow!

  xterm is in the lead!! Apparently K-nearest neighbor is the best thing to do with 100k ratings? That surprises me. We should ask the netflix master, Lester. Lester, what say you?
• 

  Jacob Abernethy (jake)
  UC Berkeley Comp Sci
  Posted at 11/13/2009

  Wow

  Y'all are still hammering away at the 100k dataset. I'm impressed. This is like a mini-netflix prize...