Experimental Test Data (ICIQ 2012)

   

icon html

This page contains all test databases that we have used in the experiments of our research paper entitled "Key-based Blocking of Duplicates in Entity-Independent Probabilistic Data" that is published in the Proceedings of the 17th International Conference on Information Quality.

   
More experimental results on blocking x-tuples can be found here.

Experiment 1: Overall Comparison of Adaptation Strategies


 

Movie Database (standard data setting) (HSQL)
DSC1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 

 

 

Experiment 2: Robustness against a varying Dirtiness of the Source Data


 

Movie Database (HSQL)
DSH1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.72
 
 
Movie Database (HSQL)
DSF1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternative per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.77
 
 
Movie Database (HSQL)
DSE1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.81
 
 
Movie Database (standard data setting) (HSQL)
DSC1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (HSQL)
DSA1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.9
 
 
Movie Database (HSQL)
DSI1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.933
 
 

 

 

Experiment 3: Robustness against a varying Data Uncertainty


 

Movie Database (standard data setting) with maximal 10 Alternatives per x-tuple (HSQL)
DSC1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (standard data setting) with maximal 15 Alternatives per x-tuple (HSQL)
DSC1_15A [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 812,272
Maximal Number of Alternatives per X-Tuple: 15
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (standard data setting) with maximal 20 Alternatives per x-tuple (HSQL)
DSC1_20A [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 1,057,191
Maximal Number of Alternatives per X-Tuple: 20
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (standard data setting) with maximal 25 Alternatives per x-tuple (HSQL)
DSC1_25A [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 1,301,207
Maximal Number of Alternatives per X-Tuple: 25
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 

 

 

Experiment 4: Uncertain Keys First


 

Movie Database (standard data setting) with maximal 10 Alternatives per x-tuple (HSQL)
DSC1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (standard data setting) with maximal 15 Alternatives per x-tuple (HSQL)
DSC1_15A [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 812,272
Maximal Number of Alternatives per X-Tuple: 15
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (standard data setting) with maximal 20 Alternatives per x-tuple (HSQL)
DSC1_20A [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 1,057,191
Maximal Number of Alternatives per X-Tuple: 20
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (standard data setting) with maximal 25 Alternatives per x-tuple (HSQL)
DSC1_25A [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 1,301,207
Maximal Number of Alternatives per X-Tuple: 25
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 

 

 

Experiment 5: Overall Comparison using Different Blocking Techniques


 

Movie Database (HSQL)
DSH1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.72
 
 
Movie Database (HSQL)
DSF1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternative per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.77
 
 
Movie Database (HSQL)
DSE1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.81
 
 
Movie Database (standard data setting) (HSQL)
DSC1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.856
 
 
Movie Database (HSQL)
DSA1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.9
 
 
Movie Database (HSQL)
DSI1 [zip]
Characteristics:
Number of X-Tuples: 102,692
Total Number of Alternatives: 561,025
Maximal Number of Alternatives per X-Tuple: 10
Number of Duplicates: 4,380
Distribution of Clustersizes (clustersize,frequence): 2,1560; 3,232; 4,72; 5,31; 6,15; 7,11; 8,9; 9,5; 10,3; 11,2; 12,1; 13,1; 15,1
Average Similarity of True Duplicates (scored with Monge-Elkan distance): 0.933