KunyuLin / Multi-view-Datasets

Multi-view datasets for classification or clustering

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Multi-view-Datasets

Multi-view datasets for classification or clustering. 

Mfeat: image, six views

https://archive.ics.uci.edu/ml/datasets/Multiple+Features

VIS/NIR: image, two views

no url

3Sources: text, three views

http://mlg.ucd.ie/datasets/3sources.html
Notice that not all the objects are descriped in three view (only 169). 
The preprocessed views is the file as follows.
3sources_global_bbc.txt
3sources_global_guardian.txt
3sources_global_reuters.txt
3sourceLabel.txt (labels)

BBC (segment): synthetic text, two, three or four views

http://mlg.ucd.ie/datasets/segment.html
The preprocessed views are the files as follows. 
bbcview1.txt
bbcview2.txt
bbcview3.txt
bbcview4.txt
bbc_label.txt (labels)

BBCSport (segment): synthetic text, two, three or four views

http://mlg.ucd.ie/datasets/segment.html
The preprocessed views are the files as follows. 
bbcsportview1.txt
bbcsportview2.txt
bbcsportview3.txt
bbcsportview4.txt
bbcsport_label.txt (labels)

ReutersEN:text (subset of Multilingual), five views

http://lig-membres.imag.fr/grimal/data.html	
The preprocessed views are the files as follows. 
reuters_sample0_2000_EN.txt
reuters_sample0_2000_FR.txt
reuters_sample0_2000_GR.txt
reuters_sample0_2000_IT.txt
reuters_sample0_2000_SP.txt
reuters_classid.txt (labels)

Animals with Attributes:image, six views

http://attributes.kyb.tuebingen.mpg.de/
A subset of the dataset is given in following files. 
AwA_sample_10_100_cq.txt
AwA_sample_10_100_lss.txt
AwA_sample_10_100_phog.txt
AwA_sample_10_100_rgsift.txt
AwA_sample_10_100_sift.txt
AwA_sample_10_100_surf.txt
AwA_sample_10_100_classid.txt (labels)

NUS-WIDE-OBJECT: image five views

http://lms.comp.nus.edu.sg/research/NUS-WIDE.htm
A subset of the dataset is given in the folder 'low level features_some5' and the name of files are as follows.
NWO_CH_test.txt
NWO_CM55_test.txt
NWO_CORR_test.txt
NWO_EDH_test.txt
NWO_WT_test.txt
NWO_test_classid.txt (labels)

WebKB: text, two views

http://lig-membres.imag.fr/grimal/data.html	
The preprocessed views of Texas are the files in folder 'webKB_texas'. 

Movies dataset: text, two views

http://lig-membres.imag.fr/grimal/data.html
The preprocessed views are the files as follows. 
M1.txt
M2.txt
Mact.txt

Cora: text, two views

http://lig-membres.imag.fr/grimal/data.html
The preprocessed views are the files as follows. 
cora_cites.txt
cora_content.txt
cora_act.txt (labels)

CiteSeer: text, two views

http://lig-membres.imag.fr/grimal/data.html
The preprocessed views are the files as follows. 
content_cites.txt
content_content.txt
content_act.txt (labels)

Multi-view Twitter

http://mlg.ucd.ie/aggregation/index.html

About

Multi-view datasets for classification or clustering