SlideShare a Scribd company logo

A Statistical and Schema Independent Approach to Identify Equivalent Properties on Linked Data

Download as PPTX, PDF
Kalpa Gunaratna, profile picture
Kalpa Gunaratna

The document discusses a schema-independent approach to identifying equivalent properties on linked data, highlighting the importance of property alignment for effective data integration. It categorizes existing techniques for property alignment into three groups: syntactic/dictionary-based, schema-dependent, and schema-independent, with the proposed method relying on statistical equivalence analyzing owl:equivalentproperty. The evaluation of this approach includes comparison with existing techniques, demonstrating effectiveness through experiments with various linked datasets.

Technology
Related topics:
Statistical Analysis • Semantic Web • Linked Data
1 of 55
Download to read offline
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
A Statistical and Schema Independent Approach to Identify Equivalent Properties on Linked Data †Kno.e.sis Center Wright State University Dayton OH, USA ‡IBM T J Watson Research Center Yorktown Heights New York NY, USA Kalpa Gunaratna†, Krishnaprasad Thirunarayan†, Prateek Jain‡, Amit Sheth†, and Sanjaya Wijeratne† {kalpa,tkprasad,amit,sanjaya}@knoesis.org, jainpr@us.ibm.com iSemantics 2013, Graz, Austria
09/06/2013 2 Motivation Why we need property alignment and it is so important? iSemantics 2013
09/06/2013 3 Many datasets. We can query! iSemantics 2013
09/06/2013 3iSemantics 2013
09/06/2013 3 Same information in different names. Therefore, data integration for better presentation is required. iSemantics 2013
Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 4 Roadmap iSemantics 2013
Existing techniques for property alignment fall into three categories. I. Syntactic/dictionary based – Uses string manipulation techniques, external dictionaries and lexical databases like WordNet. II. Schema dependent – Uses schema information such as, domain and range, definitions. III. Schema independent – Uses instance level information for the alignment.  Our approach falls under schema independent. 09/06/2013 5 Background iSemantics 2013
Properties capture meaning of triples and hence they are complex in nature.    09/06/2013 6iSemantics 2013
Properties capture meaning of triples and hence they are complex in nature. Syntactic or dictionary based approaches analyze property names for equivalence. But in LOD, name heterogeneities exist.   09/06/2013 6iSemantics 2013
Properties capture meaning of triples and hence they are complex in nature. Syntactic or dictionary based approaches analyze property names for equivalence. But in LOD, name heterogeneities exist. Therefore, syntactic or dictionary based approaches have limited coverage in property alignment.  09/06/2013 6iSemantics 2013
Properties capture meaning of triples and hence they are complex in nature. Syntactic or dictionary based approaches analyze property names for equivalence. But in LOD, name heterogeneities exist. Therefore, syntactic or dictionary based approaches have limited coverage in property alignment. Schema dependent approaches including processing domain and range, class level tags do not capture semantics of properties well. 09/06/2013 6iSemantics 2013
Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 7 Roadmap iSemantics 2013
 Statistical Equivalence is based on analyzing owl:equivalentProperty.  owl:equivalentProperty - properties that have same property extensions.          09/06/2013 8 Statistical Equivalence of properties iSemantics 2013
 Statistical Equivalence is based on analyzing owl:equivalentProperty.  owl:equivalentProperty - properties that have same property extensions. Example 1: Property P is defined by the triples, { a P b, c P d, e P f } Property Q is defined by the triples, { a Q b, c Q d, e Q f } P and Q are owl:equivalentProperty, because they have the same extension, { {a,b}, {c,d}, {e,f} }     09/06/2013 8 Statistical Equivalence of properties iSemantics 2013
 Statistical Equivalence is based on analyzing owl:equivalentProperty.  owl:equivalentProperty - properties that have same property extensions. Example 1: Property P is defined by the triples, { a P b, c P d, e P f } Property Q is defined by the triples, { a Q b, c Q d, e Q f } P and Q are owl:equivalentProperty, because they have the same extension, { {a,b}, {c,d}, {e,f} } Example 2: Property P is defined by the triples, { a P b, c P d, e P f } Property Q is defined by the triples, { a Q b, c Q d, e Q h } Then, P and Q are not owl:equivalentProperty, because their extensions are not the same. But they provide statistical evidence in support of equivalence. 09/06/2013 8 Statistical Equivalence of properties iSemantics 2013
Intuition  Higher rate of subject-object matches in extensions leads to equivalent properties. In practice, it is hard to have exact same extensions for matching properties. Because, – Datasets are incomplete. – Same instance may be modelled differently in different datasets.  Therefore, we analyze the property extensions to identify equivalent properties between datasets.  We define the following notions. Let the statement below be true for all the definitions. S1P1O1 and S2P2O2 be two triples in Dataset D1 and D2 respectively. 09/06/2013 iSemantics 2013 9
Definition 1: Candidate Match The two properties P1 and P2 are a candidate match iff S1 𝐸𝐶𝑅∗ S2 and O1 𝐸𝐶𝑅∗ O2. We say two instances are connected by an ECR* link if there is a link path between the instances using ECR links (* is the Kleene star notation). ECR links are Entity Co-reference Relationships such as those formalized using owl:sameAs and skos:exactMatch.       09/06/2013 iSemantics 2013 10
Definition 1: Candidate Match The two properties P1 and P2 are a candidate match iff S1 𝐸𝐶𝑅∗ S2 and O1 𝐸𝐶𝑅∗ O2. We say two instances are connected by an ECR* link if there is a link path between the instances using ECR links (* is the Kleene star notation). ECR links are Entity Co-reference Relationships such as those formalized using owl:sameAs and skos:exactMatch. Example The two datasets DBpedia(d) and Freebase(f) d:Arthur Purdy Stout d:place of birth d:New York City f:Arthur Purdy Stout f:place of death f:New York City   09/06/2013 iSemantics 2013 10
Definition 1: Candidate Match The two properties P1 and P2 are a candidate match iff S1 𝐸𝐶𝑅∗ S2 and O1 𝐸𝐶𝑅∗ O2. We say two instances are connected by an ECR* link if there is a link path between the instances using ECR links (* is the Kleene star notation). ECR links are Entity Co-reference Relationships such as those formalized using owl:sameAs and skos:exactMatch. Example The two datasets DBpedia(d) and Freebase(f) d:Arthur Purdy Stout d:place of birth d:New York City f:Arthur Purdy Stout f:place of death f:New York City  The above is a candidate match, but not equivalent, because intensions are different (coincidental match).  We need further analysis to decide on equivalence. 09/06/2013 iSemantics 2013 10
Match Count μ(P1,P2) – Number of triple pairs for P1 and P2 that participate in candidate matches. μ 𝑃1, 𝑃2 = | 𝑆1 𝑃1 𝑂1 ϵ 𝐷1 | ∃ 𝑆2 𝑃2 𝑂2 ϵ 𝐷2 ˄ 𝑆1 𝐸𝐶𝑅∗ 𝑆2 ˄ 𝑂1 𝐸𝐶𝑅∗ 𝑂2 | Co-appearance Count λ(P1,P2) – Number of triple pairs for P1 and P2 that have matching subjects. λ 𝑃1, 𝑃2 = | 𝑆1 𝑃1 𝑂1 ϵ 𝐷1 | ∃ 𝑆2 𝑃2 𝑂2 ϵ 𝐷2 ˄ 𝑆1 𝐸𝐶𝑅∗ 𝑆2 | Definition 2: Statistically Equivalent Properties The pair of properties P1 and P2 are statistically equivalent to degree (α, k) iff, 𝐹 = μ(𝑃1 ,𝑃2 ) λ(𝑃1 ,𝑃2 ) ≥ 𝛼, Where, μ(P1,P2) ≥ k, and 0 ˂ α ≤ 1, k > 1 09/06/2013 iSemantics 2013 11
09/06/2013 iSemantics 2013 12 Dataset 2Dataset 1 Candidate Matching Algorithm Process
09/06/2013 iSemantics 2013 12 I1 Dataset 2Dataset 1 I1=d1:Willis_Lamb Candidate Matching Algorithm Process
09/06/2013 iSemantics 2013 12 I1 I2 owl:sameAs Dataset 2Dataset 1 I1=d1:Willis_Lamb I2 =d2:willis_lamb Step 1 Candidate Matching Algorithm Process
09/06/2013 iSemantics 2013 12 I1 I2 I2 owl:sameAs P1=d1:doctoralStudent P2=d2:education. academic.advisees Dataset 2Dataset 1 d2:theodore_harold_maiman I1=d1:Willis_Lamb I2 =d2:willis_lamb I1 I1 d1:Theodore_Maiman triple 1 triple 2 triple 3 triple 4 triple 5 Step 1 Step2 Step2 Candidate Matching Algorithm Process
09/06/2013 iSemantics 2013 12 I1 I2 I2 matching resources owl:sameAs P1=d1:doctoralStudent P2=d2:education. academic.advisees Dataset 2Dataset 1 property P1 and property P2 are a candidate match d2:theodore_harold_maiman I1=d1:Willis_Lamb I2 =d2:willis_lamb I1 I1 d1:Theodore_Maiman triple 1 triple 2 triple 3 triple 4 triple 5 Step 1 Step2 Step2 Step 3 Candidate Matching Algorithm Process
09/06/2013 iSemantics 2013 13 Complexity: If the average number of properties for an entity is x and for each property, average number of objects is j. For n subjects, it requires n*j2*x2+2n comparisons. Since n > j, n > x, and x and j are independent of n, O(n).
Example: 09/06/2013 iSemantics 2013 14
Parallel computation (Map-Reduce implementation)  Generating candidate matches can be done for each instance independently. Hence, we implemented the algorithm in Hadoop 1.0.3 framework.  Generating candidate matches for instances is distributed among mappers and each mapper outputs μ and λ to the reducer for property pairs. • Map Phase – Let the number of subject instances in dataset D1 be X and namespace of dataset D2 be ns. For each subject i ϵ X, start a mapper job for GenerateCandidateMatches(i, ns). – Each mapper outputs (key,value) pairs as (p:q, μ(p,q):λ(p,q)). pϵD1 and qϵD2.  The reducer collects all μ and λ values and aggregate them for final analysis. • Reduce phase – Collects output from mappers and aggregates μ(p,q) and λ(p,q) for each key p:q.  The map reduce version on a 14 node cluster was able to achieve a speed up of 833% compared to the desktop version. 09/06/2013 iSemantics 2013 28
Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 16 Roadmap iSemantics 2013
Objectives of the evaluation – Show the effectiveness of the approach in linked datasets – Compare with existing aligning techniques   09/06/2013 iSemantics 2013 17 Evaluation
Objectives of the evaluation – Show the effectiveness of the approach in linked datasets – Compare with existing aligning techniques We selected 5000 instance samples from DBpedia, Freebase, LinkedMDB, DBLP L3S , and DBLP RKB Explorer datasets.  09/06/2013 iSemantics 2013 17 Evaluation
Objectives of the evaluation – Show the effectiveness of the approach in linked datasets – Compare with existing aligning techniques We selected 5000 instance samples from DBpedia, Freebase, LinkedMDB, DBLP L3S , and DBLP RKB Explorer datasets. These datasets have, – Complete data for instances in different viewpoints – Many inter-links – Complex properties 09/06/2013 iSemantics 2013 17 Evaluation
Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. 09/06/2013 iSemantics 2013 18
Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. 09/06/2013 iSemantics 2013 18
Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. – k can be estimated using the data as follows, – Set α = 0.5 and k = 2 (lowest positive values). – Get exact matching property (property names) pairs not identified by the algorithm and their μ – Get the average of those μ values 09/06/2013 iSemantics 2013 18
Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. – k can be estimated using the data as follows, – Set α = 0.5 and k = 2 (lowest positive values). – Get exact matching property (property names) pairs not identified by the algorithm and their μ – Get the average of those μ values – 0.92 for string similarity algorithms. 09/06/2013 iSemantics 2013 18
Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. – k can be estimated using the data as follows, – Set α = 0.5 and k = 2 (lowest positive values). – Get exact matching property (property names) pairs not identified by the algorithm and their μ – Get the average of those μ values – 0.92 for string similarity algorithms. – 0.8 for WordNet similarity. 09/06/2013 iSemantics 2013 18
Measure type DBpedia – Freebase (Person) DBpedia – Freebase (Film) DBpedia – Freebase (Software) DBpedia – LinkedMDB (Film) DBLP_RKB – DBLP_L3S (Article) Average Extension Based Algorithm Precision 0.8758 0.9737 0.6478 0.7560 1.0000 0.8427 Recall 0.8089* 0.5138 0.4339 0.8157 1.0000 0.7145 F measure 0.8410* 0.6727 0.5197 0.7848 1.0000 0.7656 WordNet Similarity Precision 0.5200 0.8620 0.7619 0.8823 1.0000 0.8052 Recall 0.4140* 0.3472 0.3018 0.3947 0.3333 0.3582 F measure 0.4609* 0.4950 0.4324 0.5454 0.5000 0.4867 Dice Similarity Precision 0.8064 0.9666 0.7659 1.0000 0.0000 0.7078 Recall 0.4777* 0.4027 0.3396 0.3421 0.0000 0.3124 F measure 0.6000* 0.5686 0.4705 0.5098 0.0000 0.4298 Jaro Similarity Precision 0.6774 0.8809 0.7755 0.9411 0.0000 0.6550 Recall 0.5350* 0.5138 0.3584 0.4210 0.0000 0.3656 F measure 0.5978* 0.6491 0.4903 0.5818 0.0000 0.4638 09/06/2013 iSemantics 2013 19 Alignment results * Marks estimated values for experiment 1 because of very large comparisons to check manually. Boldface marks highest result for each experiment.
Example identifications 09/06/2013 iSemantics 2013 20 Property pair types Dataset 1 (DBpedia) Dataset 2 (Freebase) Simple string similarity matches db:nationality fb:nationality db:religion fb:religion Synonymous matches db:occupation fb:profession db:battles fb:participated_in_conflicts Complex matches db:screenplay fb:written_by db:doctoralStudent fb:advisees
Example identifications 09/06/2013 iSemantics 2013 20 Property pair types Dataset 1 (DBpedia) Dataset 2 (Freebase) Simple string similarity matches db:nationality fb:nationality db:religion fb:religion Synonymous matches db:occupation fb:profession db:battles fb:participated_in_conflicts Complex matches db:screenplay fb:written_by db:doctoralStudent fb:advisees WordNet similarity failed to identify any of these
Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 21 Roadmap iSemantics 2013
 Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.     09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
 Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.    09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
 Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.  Some properties that are identified are intentionally different, e.g., db:distributor vs fb:production_companies. – This is because many companies produce and also distribute their films.   09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
 Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.  Some properties that are identified are intentionally different, e.g., db:distributor vs fb:production_companies. – This is because many companies produce and also distribute their films.  Some identified pairs are incorrect due to errors in data modeling. – For example, db:issue and fb:children.  09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
 Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.  Some properties that are identified are intentionally different, e.g., db:distributor vs fb:production_companies. – This is because many companies produce and also distribute their films.  Some identified pairs are incorrect due to errors in data modeling. – For example, db:issue and fb:children.  owl:sameAs linking issues in LOD (not linking exact same thing), e.g., linking London and Greater London. – We believe few misused links wont affect the algorithm as it decides on a match after analyzing many matches for a pair. 09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
 Less number of interlinks. – Evolve over time. – Look for possible other types of ECR links (i.e., rdf:seeAlso).   09/06/2013 iSemantics 2013 23
 Less number of interlinks. – Evolve over time. – Look for possible other types of ECR links (i.e., rdf:seeAlso).  Properties do not have uniform distribution in a dataset. – Hence, some properties do not have enough matches or appearances. – This is due to rare classes and domains they belong to. – We can run the algorithm on instances that these less frequent properties appear iteratively.  09/06/2013 iSemantics 2013 23
 Less number of interlinks. – Evolve over time. – Look for possible other types of ECR links (i.e., rdf:seeAlso).  Properties do not have uniform distribution in a dataset. – Hence, some properties do not have enough matches or appearances. – This is due to rare classes and domains they belong to. – We can run the algorithm on instances that these less frequent properties appear iteratively. Current limitations, – Requires ECR links – Requires overlapping datasets – Object-type properties – Inability to identify property – sub property relationships 09/06/2013 iSemantics 2013 23
Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 24 Roadmap iSemantics 2013
We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent.    09/06/2013 iSemantics 2013 25 Conclusion
We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent. This novel extension based approach works well with interlinked datasets.   09/06/2013 iSemantics 2013 25 Conclusion
We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent. This novel extension based approach works well with interlinked datasets. The extension based approach outperforms syntax or dictionary based approaches. F measure gain in the range of 57% - 78%.  09/06/2013 iSemantics 2013 25 Conclusion
We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent. This novel extension based approach works well with interlinked datasets. The extension based approach outperforms syntax or dictionary based approaches. F measure gain in the range of 57% - 78%. It requires many comparisons, but can be easily parallelized evidenced by our Map-Reduce implementation. 09/06/2013 iSemantics 2013 25 Conclusion
26 Thank You http://knoesis.wright.edu/researchers/kalpa kalpa@knoesis.org Kno.e.sis – Ohio Center of Excellence in Knowledge-enabled Computing Wright State University, Dayton, Ohio, USA Questions ? 09/06/2013 iSemantics 2013

More Related Content

PPTX
Property Alignment on Linked Open Data
Artificial Intelligence Institute at UofSC
 
PPT
R-programming-training-in-mumbai
Unmesh Baile
 
PDF
Query Optimization - Brandon Latronica
"FENG "GEORGE"" YU
 
PDF
R Programming For Beginners | R Language Tutorial | R Tutorial For Beginners ...
Edureka!
 
PPTX
Mapping Graph Queries to PostgreSQL
Gábor Szárnyas
 
PPTX
Neural Learning to Rank
Bhaskar Mitra
 
PPT
Query optimization
Neha Behl
 
PDF
record_linking
Robert Berry
 
Property Alignment on Linked Open Data
Artificial Intelligence Institute at UofSC
 
R-programming-training-in-mumbai
Unmesh Baile
 
Query Optimization - Brandon Latronica
"FENG "GEORGE"" YU
 
R Programming For Beginners | R Language Tutorial | R Tutorial For Beginners ...
Edureka!
 
Mapping Graph Queries to PostgreSQL
Gábor Szárnyas
 
Neural Learning to Rank
Bhaskar Mitra
 
Query optimization
Neha Behl
 
record_linking
Robert Berry
 

What's hot (20)

PPTX
Datastructures using c++
Gopi Nath
 
PPS
Data Structure
sheraz1
 
PDF
UNIT I LINEAR DATA STRUCTURES – LIST
Kathirvel Ayyaswamy
 
PDF
Data structure using c++
Prof. Dr. K. Adisesha
 
PPT
Data structures using C
Prof. Dr. K. Adisesha
 
PPTX
Data Structure
Karthikeyan A K
 
PPTX
Rattle Graphical Interface for R Language
Majid Abdollahi
 
PDF
Incomplete Information in RDF
Charalampos (Babis) Nikolaou
 
PDF
Exchanging more than Complete Data
net2-project
 
PPTX
2. Linear Data Structure Using Arrays - Data Structures using C++ by Varsha P...
widespreadpromotion
 
PDF
Introduction to Data Structure
Prof Ansari
 
PPT
Introduction of data structure
eShikshak
 
PDF
Data Structures (BE)
PRABHAHARAN429
 
PDF
Data Structures
Prof. Dr. K. Adisesha
 
PDF
What's next in Julia
Jiahao Chen
 
PDF
R training2
Hellen Gakuruh
 
PPT
Data structures using c
Prof. Dr. K. Adisesha
 
PDF
Elementary data structure
Biswajit Mandal
 
PDF
Introduction to R Graphics with ggplot2
izahn
 
PDF
Data Structures Notes 2021
Sreedhar Chowdam
 
Datastructures using c++
Gopi Nath
 
Data Structure
sheraz1
 
UNIT I LINEAR DATA STRUCTURES – LIST
Kathirvel Ayyaswamy
 
Data structure using c++
Prof. Dr. K. Adisesha
 
Data structures using C
Prof. Dr. K. Adisesha
 
Data Structure
Karthikeyan A K
 
Rattle Graphical Interface for R Language
Majid Abdollahi
 
Incomplete Information in RDF
Charalampos (Babis) Nikolaou
 
Exchanging more than Complete Data
net2-project
 
2. Linear Data Structure Using Arrays - Data Structures using C++ by Varsha P...
widespreadpromotion
 
Introduction to Data Structure
Prof Ansari
 
Introduction of data structure
eShikshak
 
Data Structures (BE)
PRABHAHARAN429
 
Data Structures
Prof. Dr. K. Adisesha
 
What's next in Julia
Jiahao Chen
 
R training2
Hellen Gakuruh
 
Data structures using c
Prof. Dr. K. Adisesha
 
Elementary data structure
Biswajit Mandal
 
Introduction to R Graphics with ggplot2
izahn
 
Data Structures Notes 2021
Sreedhar Chowdam
 
Ad

Similar to A Statistical and Schema Independent Approach to Identify Equivalent Properties on Linked Data (20)

PPT
Data Integration Ontology Mapping
Pradeep B Pillai
 
PDF
Information Retrieval using Semantic Similarity
Saswat Padhi
 
PDF
FEATURES MATCHING USING NATURAL LANGUAGE PROCESSING
IJCI JOURNAL
 
PDF
[Evaldas Taroza - Master thesis] Schema Matching and Automatic Web Data Extra...
Evaldas Taroza
 
PDF
Conceptual similarity measurement algorithm for domain specific ontology[
Zac Darcy
 
PDF
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
Zac Darcy
 
PDF
CONCEPTUAL SIMILARITY MEASUREMENT ALGORITHM FOR DOMAIN SPECIFIC ONTOLOGY
Zac Darcy
 
PDF
On the Impact of sameAs on Schema Matching
Joe Raad
 
PPTX
Towards a Distributional Semantic Web Stack
Andre Freitas
 
PPT
Identifying Value Mappings for Data Integration_PVERConf_May2011
NORC at the University of Chicago
 
PPTX
Logical Detection of Invalid SameAs Statements in RDF Data
University PARIS-SUD
 
PDF
Semantic Analysis Using MapReduce
Ankur Pandey
 
PDF
paper
Tengyuan Ye
 
PDF
Document Retrieval System, a Case Study
IJERA Editor
 
PDF
Hybrid approach for generating non overlapped substring using genetic algorithm
eSAT Publishing House
 
PDF
Measure Term Similarity Using a Semantic Network Approach
BOHR International Journal of Intelligent Instrumentation and Computing
 
PDF
Exploiting Distributional Semantic Models in Question Answering
Pierpaolo Basile
 
PDF
Pay-as-you-go Reconciliation in Schema Matching Networks
PlanetData Network of Excellence
 
PDF
A Primer on Entity Resolution
Benjamin Bengfort
 
PDF
Semantic Similarity Measures for Semantic Relation Extraction
Alexander Panchenko
 
Data Integration Ontology Mapping
Pradeep B Pillai
 
Information Retrieval using Semantic Similarity
Saswat Padhi
 
FEATURES MATCHING USING NATURAL LANGUAGE PROCESSING
IJCI JOURNAL
 
[Evaldas Taroza - Master thesis] Schema Matching and Automatic Web Data Extra...
Evaldas Taroza
 
Conceptual similarity measurement algorithm for domain specific ontology[
Zac Darcy
 
Conceptual Similarity Measurement Algorithm For Domain Specific Ontology
Zac Darcy
 
CONCEPTUAL SIMILARITY MEASUREMENT ALGORITHM FOR DOMAIN SPECIFIC ONTOLOGY
Zac Darcy
 
On the Impact of sameAs on Schema Matching
Joe Raad
 
Towards a Distributional Semantic Web Stack
Andre Freitas
 
Identifying Value Mappings for Data Integration_PVERConf_May2011
NORC at the University of Chicago
 
Logical Detection of Invalid SameAs Statements in RDF Data
University PARIS-SUD
 
Semantic Analysis Using MapReduce
Ankur Pandey
 
paper
Tengyuan Ye
 
Document Retrieval System, a Case Study
IJERA Editor
 
Hybrid approach for generating non overlapped substring using genetic algorithm
eSAT Publishing House
 
Measure Term Similarity Using a Semantic Network Approach
BOHR International Journal of Intelligent Instrumentation and Computing
 
Exploiting Distributional Semantic Models in Question Answering
Pierpaolo Basile
 
Pay-as-you-go Reconciliation in Schema Matching Networks
PlanetData Network of Excellence
 
A Primer on Entity Resolution
Benjamin Bengfort
 
Semantic Similarity Measures for Semantic Relation Extraction
Alexander Panchenko
 
Ad

Recently uploaded (20)

PDF
NewMind AI Weekly Chronicles - August'25 Week I
NewMind AI
 
PDF
Spectral efficient network and resource selection model in 5G networks
IAESIJAI
 
PDF
Network Security Unit 5.pdf for BCA BBA.
Serpent6
 
PDF
The Rise and Fall of 3GPP – Time for a Sabbatical?
3G4G
 
PPTX
sap open course for s4hana steps from ECC to s4
sreeni2106invites
 
PPTX
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
ReZa AdineH
 
PPTX
MYSQL Presentation for SQL database connectivity
Swati270511
 
PDF
Dropbox Q2 2025 Financial Results & Investor Presentation
Dropbox
 
PDF
Optimiser vos workloads AI/ML sur Amazon EC2 et AWS Graviton
Julien SIMON
 
PPTX
Spectroscopy.pptx food analysis technology
nawahassan45
 
PDF
cuic standard and advanced reporting.pdf
SimoneTitaniumTomase
 
PPTX
20250228 LYD VKU AI Blended-Learning.pptx
DatVoNgoc
 
PDF
Encapsulation_ Review paper, used for researhc scholars
gurumoop
 
PDF
Profit Center Accounting in SAP S/4HANA, S4F28 Col11
Libreria ERP
 
PDF
Per capita expenditure prediction using model stacking based on satellite ima...
IAESIJAI
 
PDF
Unlocking AI with Model Context Protocol (MCP)
Brian McKeiver
 
PDF
Approach and Philosophy of On baking technology
30anthuong17
 
PDF
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
Safe Software
 
PDF
Encapsulation theory and applications.pdf
gurumoop
 
PDF
KodekX | Application Modernization Development
KodekX
 
NewMind AI Weekly Chronicles - August'25 Week I
NewMind AI
 
Spectral efficient network and resource selection model in 5G networks
IAESIJAI
 
Network Security Unit 5.pdf for BCA BBA.
Serpent6
 
The Rise and Fall of 3GPP – Time for a Sabbatical?
3G4G
 
sap open course for s4hana steps from ECC to s4
sreeni2106invites
 
Detection-First SIEM: Rule Types, Dashboards, and Threat-Informed Strategy
ReZa AdineH
 
MYSQL Presentation for SQL database connectivity
Swati270511
 
Dropbox Q2 2025 Financial Results & Investor Presentation
Dropbox
 
Optimiser vos workloads AI/ML sur Amazon EC2 et AWS Graviton
Julien SIMON
 
Spectroscopy.pptx food analysis technology
nawahassan45
 
cuic standard and advanced reporting.pdf
SimoneTitaniumTomase
 
20250228 LYD VKU AI Blended-Learning.pptx
DatVoNgoc
 
Encapsulation_ Review paper, used for researhc scholars
gurumoop
 
Profit Center Accounting in SAP S/4HANA, S4F28 Col11
Libreria ERP
 
Per capita expenditure prediction using model stacking based on satellite ima...
IAESIJAI
 
Unlocking AI with Model Context Protocol (MCP)
Brian McKeiver
 
Approach and Philosophy of On baking technology
30anthuong17
 
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
Safe Software
 
Encapsulation theory and applications.pdf
gurumoop
 
KodekX | Application Modernization Development
KodekX
 

A Statistical and Schema Independent Approach to Identify Equivalent Properties on Linked Data

  • 1. A Statistical and Schema Independent Approach to Identify Equivalent Properties on Linked Data †Kno.e.sis Center Wright State University Dayton OH, USA ‡IBM T J Watson Research Center Yorktown Heights New York NY, USA Kalpa Gunaratna†, Krishnaprasad Thirunarayan†, Prateek Jain‡, Amit Sheth†, and Sanjaya Wijeratne† {kalpa,tkprasad,amit,sanjaya}@knoesis.org, jainpr@us.ibm.com iSemantics 2013, Graz, Austria
  • 2. 09/06/2013 2 Motivation Why we need property alignment and it is so important? iSemantics 2013
  • 3. 09/06/2013 3 Many datasets. We can query! iSemantics 2013
  • 5. 09/06/2013 3 Same information in different names. Therefore, data integration for better presentation is required. iSemantics 2013
  • 6. Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 4 Roadmap iSemantics 2013
  • 7. Existing techniques for property alignment fall into three categories. I. Syntactic/dictionary based – Uses string manipulation techniques, external dictionaries and lexical databases like WordNet. II. Schema dependent – Uses schema information such as, domain and range, definitions. III. Schema independent – Uses instance level information for the alignment.  Our approach falls under schema independent. 09/06/2013 5 Background iSemantics 2013
  • 8. Properties capture meaning of triples and hence they are complex in nature.    09/06/2013 6iSemantics 2013
  • 9. Properties capture meaning of triples and hence they are complex in nature. Syntactic or dictionary based approaches analyze property names for equivalence. But in LOD, name heterogeneities exist.   09/06/2013 6iSemantics 2013
  • 10. Properties capture meaning of triples and hence they are complex in nature. Syntactic or dictionary based approaches analyze property names for equivalence. But in LOD, name heterogeneities exist. Therefore, syntactic or dictionary based approaches have limited coverage in property alignment.  09/06/2013 6iSemantics 2013
  • 11. Properties capture meaning of triples and hence they are complex in nature. Syntactic or dictionary based approaches analyze property names for equivalence. But in LOD, name heterogeneities exist. Therefore, syntactic or dictionary based approaches have limited coverage in property alignment. Schema dependent approaches including processing domain and range, class level tags do not capture semantics of properties well. 09/06/2013 6iSemantics 2013
  • 12. Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 7 Roadmap iSemantics 2013
  • 13.  Statistical Equivalence is based on analyzing owl:equivalentProperty.  owl:equivalentProperty - properties that have same property extensions.          09/06/2013 8 Statistical Equivalence of properties iSemantics 2013
  • 14.  Statistical Equivalence is based on analyzing owl:equivalentProperty.  owl:equivalentProperty - properties that have same property extensions. Example 1: Property P is defined by the triples, { a P b, c P d, e P f } Property Q is defined by the triples, { a Q b, c Q d, e Q f } P and Q are owl:equivalentProperty, because they have the same extension, { {a,b}, {c,d}, {e,f} }     09/06/2013 8 Statistical Equivalence of properties iSemantics 2013
  • 15.  Statistical Equivalence is based on analyzing owl:equivalentProperty.  owl:equivalentProperty - properties that have same property extensions. Example 1: Property P is defined by the triples, { a P b, c P d, e P f } Property Q is defined by the triples, { a Q b, c Q d, e Q f } P and Q are owl:equivalentProperty, because they have the same extension, { {a,b}, {c,d}, {e,f} } Example 2: Property P is defined by the triples, { a P b, c P d, e P f } Property Q is defined by the triples, { a Q b, c Q d, e Q h } Then, P and Q are not owl:equivalentProperty, because their extensions are not the same. But they provide statistical evidence in support of equivalence. 09/06/2013 8 Statistical Equivalence of properties iSemantics 2013
  • 16. Intuition  Higher rate of subject-object matches in extensions leads to equivalent properties. In practice, it is hard to have exact same extensions for matching properties. Because, – Datasets are incomplete. – Same instance may be modelled differently in different datasets.  Therefore, we analyze the property extensions to identify equivalent properties between datasets.  We define the following notions. Let the statement below be true for all the definitions. S1P1O1 and S2P2O2 be two triples in Dataset D1 and D2 respectively. 09/06/2013 iSemantics 2013 9
  • 17. Definition 1: Candidate Match The two properties P1 and P2 are a candidate match iff S1 𝐸𝐶𝑅∗ S2 and O1 𝐸𝐶𝑅∗ O2. We say two instances are connected by an ECR* link if there is a link path between the instances using ECR links (* is the Kleene star notation). ECR links are Entity Co-reference Relationships such as those formalized using owl:sameAs and skos:exactMatch.       09/06/2013 iSemantics 2013 10
  • 18. Definition 1: Candidate Match The two properties P1 and P2 are a candidate match iff S1 𝐸𝐶𝑅∗ S2 and O1 𝐸𝐶𝑅∗ O2. We say two instances are connected by an ECR* link if there is a link path between the instances using ECR links (* is the Kleene star notation). ECR links are Entity Co-reference Relationships such as those formalized using owl:sameAs and skos:exactMatch. Example The two datasets DBpedia(d) and Freebase(f) d:Arthur Purdy Stout d:place of birth d:New York City f:Arthur Purdy Stout f:place of death f:New York City   09/06/2013 iSemantics 2013 10
  • 19. Definition 1: Candidate Match The two properties P1 and P2 are a candidate match iff S1 𝐸𝐶𝑅∗ S2 and O1 𝐸𝐶𝑅∗ O2. We say two instances are connected by an ECR* link if there is a link path between the instances using ECR links (* is the Kleene star notation). ECR links are Entity Co-reference Relationships such as those formalized using owl:sameAs and skos:exactMatch. Example The two datasets DBpedia(d) and Freebase(f) d:Arthur Purdy Stout d:place of birth d:New York City f:Arthur Purdy Stout f:place of death f:New York City  The above is a candidate match, but not equivalent, because intensions are different (coincidental match).  We need further analysis to decide on equivalence. 09/06/2013 iSemantics 2013 10
  • 20. Match Count μ(P1,P2) – Number of triple pairs for P1 and P2 that participate in candidate matches. μ 𝑃1, 𝑃2 = | 𝑆1 𝑃1 𝑂1 ϵ 𝐷1 | ∃ 𝑆2 𝑃2 𝑂2 ϵ 𝐷2 ˄ 𝑆1 𝐸𝐶𝑅∗ 𝑆2 ˄ 𝑂1 𝐸𝐶𝑅∗ 𝑂2 | Co-appearance Count λ(P1,P2) – Number of triple pairs for P1 and P2 that have matching subjects. λ 𝑃1, 𝑃2 = | 𝑆1 𝑃1 𝑂1 ϵ 𝐷1 | ∃ 𝑆2 𝑃2 𝑂2 ϵ 𝐷2 ˄ 𝑆1 𝐸𝐶𝑅∗ 𝑆2 | Definition 2: Statistically Equivalent Properties The pair of properties P1 and P2 are statistically equivalent to degree (α, k) iff, 𝐹 = μ(𝑃1 ,𝑃2 ) λ(𝑃1 ,𝑃2 ) ≥ 𝛼, Where, μ(P1,P2) ≥ k, and 0 ˂ α ≤ 1, k > 1 09/06/2013 iSemantics 2013 11
  • 21. 09/06/2013 iSemantics 2013 12 Dataset 2Dataset 1 Candidate Matching Algorithm Process
  • 22. 09/06/2013 iSemantics 2013 12 I1 Dataset 2Dataset 1 I1=d1:Willis_Lamb Candidate Matching Algorithm Process
  • 23. 09/06/2013 iSemantics 2013 12 I1 I2 owl:sameAs Dataset 2Dataset 1 I1=d1:Willis_Lamb I2 =d2:willis_lamb Step 1 Candidate Matching Algorithm Process
  • 24. 09/06/2013 iSemantics 2013 12 I1 I2 I2 owl:sameAs P1=d1:doctoralStudent P2=d2:education. academic.advisees Dataset 2Dataset 1 d2:theodore_harold_maiman I1=d1:Willis_Lamb I2 =d2:willis_lamb I1 I1 d1:Theodore_Maiman triple 1 triple 2 triple 3 triple 4 triple 5 Step 1 Step2 Step2 Candidate Matching Algorithm Process
  • 25. 09/06/2013 iSemantics 2013 12 I1 I2 I2 matching resources owl:sameAs P1=d1:doctoralStudent P2=d2:education. academic.advisees Dataset 2Dataset 1 property P1 and property P2 are a candidate match d2:theodore_harold_maiman I1=d1:Willis_Lamb I2 =d2:willis_lamb I1 I1 d1:Theodore_Maiman triple 1 triple 2 triple 3 triple 4 triple 5 Step 1 Step2 Step2 Step 3 Candidate Matching Algorithm Process
  • 26. 09/06/2013 iSemantics 2013 13 Complexity: If the average number of properties for an entity is x and for each property, average number of objects is j. For n subjects, it requires n*j2*x2+2n comparisons. Since n > j, n > x, and x and j are independent of n, O(n).
  • 28. Parallel computation (Map-Reduce implementation)  Generating candidate matches can be done for each instance independently. Hence, we implemented the algorithm in Hadoop 1.0.3 framework.  Generating candidate matches for instances is distributed among mappers and each mapper outputs μ and λ to the reducer for property pairs. • Map Phase – Let the number of subject instances in dataset D1 be X and namespace of dataset D2 be ns. For each subject i ϵ X, start a mapper job for GenerateCandidateMatches(i, ns). – Each mapper outputs (key,value) pairs as (p:q, μ(p,q):λ(p,q)). pϵD1 and qϵD2.  The reducer collects all μ and λ values and aggregate them for final analysis. • Reduce phase – Collects output from mappers and aggregates μ(p,q) and λ(p,q) for each key p:q.  The map reduce version on a 14 node cluster was able to achieve a speed up of 833% compared to the desktop version. 09/06/2013 iSemantics 2013 28
  • 29. Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 16 Roadmap iSemantics 2013
  • 30. Objectives of the evaluation – Show the effectiveness of the approach in linked datasets – Compare with existing aligning techniques   09/06/2013 iSemantics 2013 17 Evaluation
  • 31. Objectives of the evaluation – Show the effectiveness of the approach in linked datasets – Compare with existing aligning techniques We selected 5000 instance samples from DBpedia, Freebase, LinkedMDB, DBLP L3S , and DBLP RKB Explorer datasets.  09/06/2013 iSemantics 2013 17 Evaluation
  • 32. Objectives of the evaluation – Show the effectiveness of the approach in linked datasets – Compare with existing aligning techniques We selected 5000 instance samples from DBpedia, Freebase, LinkedMDB, DBLP L3S , and DBLP RKB Explorer datasets. These datasets have, – Complete data for instances in different viewpoints – Many inter-links – Complex properties 09/06/2013 iSemantics 2013 17 Evaluation
  • 33. Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. 09/06/2013 iSemantics 2013 18
  • 34. Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. 09/06/2013 iSemantics 2013 18
  • 35. Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. – k can be estimated using the data as follows, – Set α = 0.5 and k = 2 (lowest positive values). – Get exact matching property (property names) pairs not identified by the algorithm and their μ – Get the average of those μ values 09/06/2013 iSemantics 2013 18
  • 36. Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. – k can be estimated using the data as follows, – Set α = 0.5 and k = 2 (lowest positive values). – Get exact matching property (property names) pairs not identified by the algorithm and their μ – Get the average of those μ values – 0.92 for string similarity algorithms. 09/06/2013 iSemantics 2013 18
  • 37. Experiment details – α = 0.5 for all experiments (works for LOD) except DBpedia and Freebase movie alignment where it was 0.7. – k was set as 14, 6, 2, 2, and 2 respectively for Person, Film and Software between DBpedia and Freebase, Film between LinkedMDB and DBpedia, and article between DBLP datasets. – k can be estimated using the data as follows, – Set α = 0.5 and k = 2 (lowest positive values). – Get exact matching property (property names) pairs not identified by the algorithm and their μ – Get the average of those μ values – 0.92 for string similarity algorithms. – 0.8 for WordNet similarity. 09/06/2013 iSemantics 2013 18
  • 38. Measure type DBpedia – Freebase (Person) DBpedia – Freebase (Film) DBpedia – Freebase (Software) DBpedia – LinkedMDB (Film) DBLP_RKB – DBLP_L3S (Article) Average Extension Based Algorithm Precision 0.8758 0.9737 0.6478 0.7560 1.0000 0.8427 Recall 0.8089* 0.5138 0.4339 0.8157 1.0000 0.7145 F measure 0.8410* 0.6727 0.5197 0.7848 1.0000 0.7656 WordNet Similarity Precision 0.5200 0.8620 0.7619 0.8823 1.0000 0.8052 Recall 0.4140* 0.3472 0.3018 0.3947 0.3333 0.3582 F measure 0.4609* 0.4950 0.4324 0.5454 0.5000 0.4867 Dice Similarity Precision 0.8064 0.9666 0.7659 1.0000 0.0000 0.7078 Recall 0.4777* 0.4027 0.3396 0.3421 0.0000 0.3124 F measure 0.6000* 0.5686 0.4705 0.5098 0.0000 0.4298 Jaro Similarity Precision 0.6774 0.8809 0.7755 0.9411 0.0000 0.6550 Recall 0.5350* 0.5138 0.3584 0.4210 0.0000 0.3656 F measure 0.5978* 0.6491 0.4903 0.5818 0.0000 0.4638 09/06/2013 iSemantics 2013 19 Alignment results * Marks estimated values for experiment 1 because of very large comparisons to check manually. Boldface marks highest result for each experiment.
  • 39. Example identifications 09/06/2013 iSemantics 2013 20 Property pair types Dataset 1 (DBpedia) Dataset 2 (Freebase) Simple string similarity matches db:nationality fb:nationality db:religion fb:religion Synonymous matches db:occupation fb:profession db:battles fb:participated_in_conflicts Complex matches db:screenplay fb:written_by db:doctoralStudent fb:advisees
  • 40. Example identifications 09/06/2013 iSemantics 2013 20 Property pair types Dataset 1 (DBpedia) Dataset 2 (Freebase) Simple string similarity matches db:nationality fb:nationality db:religion fb:religion Synonymous matches db:occupation fb:profession db:battles fb:participated_in_conflicts Complex matches db:screenplay fb:written_by db:doctoralStudent fb:advisees WordNet similarity failed to identify any of these
  • 41. Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 21 Roadmap iSemantics 2013
  • 42.  Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.     09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
  • 43.  Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.    09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
  • 44.  Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.  Some properties that are identified are intentionally different, e.g., db:distributor vs fb:production_companies. – This is because many companies produce and also distribute their films.   09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
  • 45.  Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.  Some properties that are identified are intentionally different, e.g., db:distributor vs fb:production_companies. – This is because many companies produce and also distribute their films.  Some identified pairs are incorrect due to errors in data modeling. – For example, db:issue and fb:children.  09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
  • 46.  Our experiment covered multi-domain to multi-domain, multi- domain to specific domain and specific-domain to specific-domain dataset property alignment.  In every experiment, the extension based algorithm outperformed others (F measure). F measure gain is in the range of 57% to 78%.  Some properties that are identified are intentionally different, e.g., db:distributor vs fb:production_companies. – This is because many companies produce and also distribute their films.  Some identified pairs are incorrect due to errors in data modeling. – For example, db:issue and fb:children.  owl:sameAs linking issues in LOD (not linking exact same thing), e.g., linking London and Greater London. – We believe few misused links wont affect the algorithm as it decides on a match after analyzing many matches for a pair. 09/06/2013 iSemantics 2013 22 Discussion, interesting facts, and future directions
  • 47.  Less number of interlinks. – Evolve over time. – Look for possible other types of ECR links (i.e., rdf:seeAlso).   09/06/2013 iSemantics 2013 23
  • 48.  Less number of interlinks. – Evolve over time. – Look for possible other types of ECR links (i.e., rdf:seeAlso).  Properties do not have uniform distribution in a dataset. – Hence, some properties do not have enough matches or appearances. – This is due to rare classes and domains they belong to. – We can run the algorithm on instances that these less frequent properties appear iteratively.  09/06/2013 iSemantics 2013 23
  • 49.  Less number of interlinks. – Evolve over time. – Look for possible other types of ECR links (i.e., rdf:seeAlso).  Properties do not have uniform distribution in a dataset. – Hence, some properties do not have enough matches or appearances. – This is due to rare classes and domains they belong to. – We can run the algorithm on instances that these less frequent properties appear iteratively. Current limitations, – Requires ECR links – Requires overlapping datasets – Object-type properties – Inability to identify property – sub property relationships 09/06/2013 iSemantics 2013 23
  • 50. Background Statistical Equivalence of properties Evaluation Discussion, interesting facts, and future directions Conclusion 09/06/2013 24 Roadmap iSemantics 2013
  • 51. We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent.    09/06/2013 iSemantics 2013 25 Conclusion
  • 52. We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent. This novel extension based approach works well with interlinked datasets.   09/06/2013 iSemantics 2013 25 Conclusion
  • 53. We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent. This novel extension based approach works well with interlinked datasets. The extension based approach outperforms syntax or dictionary based approaches. F measure gain in the range of 57% - 78%.  09/06/2013 iSemantics 2013 25 Conclusion
  • 54. We approximate owl:equivalentProperty using Statistical Equivalence of properties by analyzing property extensions, which is schema independent. This novel extension based approach works well with interlinked datasets. The extension based approach outperforms syntax or dictionary based approaches. F measure gain in the range of 57% - 78%. It requires many comparisons, but can be easily parallelized evidenced by our Map-Reduce implementation. 09/06/2013 iSemantics 2013 25 Conclusion
  • 55. 26 Thank You http://knoesis.wright.edu/researchers/kalpa kalpa@knoesis.org Kno.e.sis – Ohio Center of Excellence in Knowledge-enabled Computing Wright State University, Dayton, Ohio, USA Questions ? 09/06/2013 iSemantics 2013

Editor's Notes

  • #34: k in some datasets are low because there are low appearing property pairs as well as popular property pairs. Because of this, sometimes lower k values as 2 produces better results but may be with low confidence.
  • #35: k in some datasets are low because there are low appearing property pairs as well as popular property pairs. Because of this, sometimes lower k values as 2 produces better results but may be with low confidence.
  • #36: k in some datasets are low because there are low appearing property pairs as well as popular property pairs. Because of this, sometimes lower k values as 2 produces better results but may be with low confidence.
  • #37: k in some datasets are low because there are low appearing property pairs as well as popular property pairs. Because of this, sometimes lower k values as 2 produces better results but may be with low confidence.
  • #38: k in some datasets are low because there are low appearing property pairs as well as popular property pairs. Because of this, sometimes lower k values as 2 produces better results but may be with low confidence.