SlideShare ist ein Scribd-Unternehmen logo
1 von 26
Downloaden Sie, um offline zu lesen
Introduction Dimensions Results




          Comparison of Reasoners for large Ontologies in
                     the OWL 2 EL Profile
                                                         WAI



                                              Kathrin Dentler


                                           November 8th 2010



Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   1/26
Introduction Dimensions Results

 Outline




      1 Introduction

      2 Dimensions

      3 Results




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   2/26
Introduction Dimensions Results

 Outline




      1 Introduction

      2 Dimensions

      3 Results




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   3/26
Introduction Dimensions Results

 PhD Project
 Supervisors




                 Ronald Cornet                                Annette ten Teije




                 Nicolette de Keizer                          Frank van Harmelen

Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   4/26
Introduction Dimensions Results

 PhD Project
 SEMI4

      SEMantic Interoperability for Integration of clinical
      Information and Indicators

      Quality Indicators
      General Problem: growing number, labour intensive
      Approach:
          formalize indicators;
          integrate data sources;
          investigate which adjustments are needed to facilitate
          data-capture at the required quality levels


      → More in next WAI
Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   5/26
Introduction Dimensions Results

 SNOMED CT
 Systematized Nomenclature of Medicine, Clinical Terms




      SNOMED CT
         used to represent information in electronic health records
         around 300,000 active concepts
         available as DB tables and as OWL ontology
         in tractable OWL 2 profile EL


      Goal: Classify SNOMED CT with a reasoner to obtain the
      closure, i.e. all possible inferences.



Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   6/26
Introduction Dimensions Results

 An example OWL 2 EL ontology

      Fracture   Traumatic abnormality by morphology
      Traumatic abnormality by morphology   Traumatic abnormality
      Traumatic abnormality   Damage

      Fracture of foot ≡ Fracture of lower limb ∃.rolegroup
                         (∃associated morphology.Fracture
                           ∃finding site.Bone structure of foot)

      Fracture of lower limb ≡ Fracture of bone ∃rolegroup.
                          (∃associated morphology.Fracture
                            ∃finding site.Bone structure of lower limb)

      Inferred:
      Fracture      Damage

      Fracture of foot    Fracture of lower limb
      Pellet’s explanation:
      Fracture of foot ≡ Fracture of bone ∃.rolegroup
                          (∃associated morphology.Fracture
                            ∃finding site.Bone structure of foot)


Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   7/26
Introduction Dimensions Results

 Included Reasoners



              CB
              CEL
              FaCT++
              HermiT
              Pellet
              Racer
              Snorocket
              TrOWL



Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   8/26
Introduction Dimensions Results

 Related Work
 Earlier Comparisons of Classification Performance for SNOMED CT


                                                                10000                                         CB
                                                                                                             CEL
                                                                                                            FaCT
                                                                                                           Racer
                                                                                                        Snorocket
                   Classification Time in Seconds (log-scale)




                                                                1000




                                                                 100




                                                                  10
                                                                        2005   2006     2008     2009    2010
                                                                                      Timeline




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile                               9/26
Introduction Dimensions Results

 Outline




      1 Introduction

      2 Dimensions

      3 Results




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   10/26
Introduction Dimensions Results

 Dimensions




          1   Reasoning Characteristics
          2   Practical Usability
          3   Performance Indicators




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   11/26
Introduction Dimensions Results

 Dimension Reasoning Characteristics
 Independent of Ontology




              Methodology
              Soundness and Completeness in Theory
              Expressivity and Computational Complexity
              Incremental Classification
              Rule Support
              Justifications
              Support of ABox Reasoning Tasks




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   12/26
Introduction Dimensions Results

 Dimension Practical Usability
 Independent of Ontology



              OWL API
              OWLlink
              Availability as Prot´g´ Plugin
                                  e e
              License
              Open or Closed Source
              Programming Language
              Supported Platforms
              Jena
              Kind of Institution


Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   13/26
Introduction Dimensions Results

 Dimension Performance Indicators
 Related to Ontology




              Classification Performance
              TBox Consistency Checking Performance
              Concept Satisfiability Checking Performance
              Subsumption Query Performance
              Required Heap Space
              Soundness and Completeness in Practice




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   14/26
Introduction Dimensions Results

 Outline




      1 Introduction

      2 Dimensions

      3 Results




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   15/26
Introduction Dimensions Results

 Results: Categorization of Reasoners
 Dimension Reasoning Characteristics



                        CB          CEL      FaCT++      HermiT     Pellet       RP         SR       TrOWL
                                                                                                     (REL)
        Methodology     consequ.-   compl.   tableau-    hyper      tableau      tableau    compl.   approx.
                        based       rules    based       tableau    based        based      rules    (compl.
                                                                                                     rules)
        Sound
                                                                                                         ( )
        Complete
                                                                                                     x( )
        IC
        (addition/      x/x             /x   x/x         x/x          /          x/x            /x   x/x
        removal)
        Rule            x           x        x                                              x        x
        Support
                                                         (SWRL)     (SWRL)       (SWRL,
                                                                                 nRQL)
        Justifications   x                    x           x                                  x        x
        ABox            x                                                                   x
        Reasoning
                                                                    (SPARQL)     (SPARQL,            (SPARQL)
                                                                                 nRQL)




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile                           16/26
Introduction Dimensions Results

 Results: Categorization of Reasoners
 Dimension Practical Usability




                         CB              CEL             FaCT++      HermiT     Pellet       RP     SR     TrOWL
        OWL API          x
        OWLlink API      x                                                                          x      x
        Prot´g´ Plugin
             e e         x                                                                   x
        License          DuLi: GLGPL     AP 2.0          GLGPL       GLGPL      DuLi: AGPL   own    own    DuLi: AGPL
        Open Source                                                                          x      x      x
        Language         OCaml           Common Lisp     C++         Java       Java         Lisp   Java   Java
        Platforms        all             Linux           all         all        all          all    all    all
        Jena             x               x               x           x                       x      x      x
        Institution      a               a               a           a          c            c      g      a




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile                              17/26
Introduction Dimensions Results

 Results: Categorization of Reasoners
 Dimension Performance Indicators




      Experiments on Galjas
      With: (GO, NCI and) SNOMED CT




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   18/26
Introduction Dimensions Results

 Classification Performance vs. Supported Expressivity
 Dimension Performance Indicators

                                                                                                   Supported Expressivity
                                                                           EL       EL        EL       EL    Horn SHIQ SHIQ(D-) SROIQ(D) SROIQ(D)
                                                                              G0
                                                                             NCI
                                                                       SNOMED CT

                                                                1000
                   Classification Time in Seconds (log-scale)




                                                                100




                                                                 10




                                                                  1




                                                                 0.1
                                                                          CEL      Pellet Snorocket    TR         CB     RacerPro HermiT     FaCT++
                                                                                    Reasoners ordered by increasing Supported Expressivity




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile                                                                 19/26
Introduction Dimensions Results

 Consistency Checking Performance
 Dimension Performance Indicators




                               CB CEL FaCT++ HermiT Pellet RP SR TR
        OSNOMED CT              x 0.88    15.3  0.00 16.78 x 0.00 0.00

      CEL and TrOWL REL need to classify first, but return true
      anyway, Snorocket outputs a warning.




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   20/26
Introduction Dimensions Results

 Concept Satisfiability Checking Performance
 Dimension Performance Indicators




                          CB CEL FaCT++ HermiT Pellet       RP SR TR
        OS CT BC           x 38.42  22.37 5,276.85 56.91 273.45 0.07 5.17
        OS CT AC           x 34.59   1.76     1.36 6.07 0.00 0.06 0.46

      E.g TrOWL returns reliable results only after the ontology is
      classified.
      → benchmarks with inconsistent ontologies / unsatisfiable
      concepts needed!




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   21/26
Introduction Dimensions Results

 Subsumption Query Performance
 Dimension Performance Indicators


                        CB CEL FaCT++ HermiT                               Pellet        RP   SR TR
        NC BC
        # results          x    1              20       20       20       20   x    0
        seconds            x 0.96          701.79 6,649.85 2,793.31 3,380.67 NPE 0.17
        AC BC
        # results          x    1                 1            20              20         20   x    0
        seconds            x 0.00              0.06         16.94            0.49       0.74 NPE 0.00
        NC AC
        # results          x 20                  20             20             20         20 20 20
        seconds            x 0.00              0.00           0.00           0.00       0.70 0.00 0.28
        AC AC
        # results          x    1                 1            20              20         20 20      0
        seconds            x 0.00              0.06         17.12            0.00       0.92 6.97 0.00

Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile                    22/26
Introduction Dimensions Results

 Minimum Heap Space for Java Reasoners
 Dimension Performance Indicators




                  CB CEL FaCT++ HermiT Pellet RP SR TR
                  n/a n/a n/a   4.5    10     n/a 2.5 4




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   23/26
Introduction Dimensions Results

 Soundness and Completeness in Practice
 Dimension Performance Indicators




      CB: missing SubClassOf owl:Thing for top-level classes

      CEL: additional EquivalentClasses and EquivalentProperties
      for each class and property

      RacerPro: additional sub class owl:Nothing for all leaf concepts




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   24/26
Introduction Dimensions Results

 Soundness and Completeness in Practice
 Dimension Performance Indicators




      Table: SNOMED CT: Missing / Additional inferred SubClassOf
      statements in regard to the Relationships Table
                              CB CEL FaCT++ HermiT Pellet RP SR TR
           Missing             0   0      0      0 386 0 86 0
           Additional          0   0      0      0 546 0 34 0




Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile   25/26
Introduction Dimensions Results




      Questions ?


                                                                                        Kathrin Dentler
                                                                                k.dentler@amc.uva.nl
                                                                                  k.dentler@few.vu.nl
                                                                           http://reasonerbench.net/

Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile                     26/26

Weitere ähnliche Inhalte

Ähnlich wie Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile

Qualifications And Experience Presentation
Qualifications And Experience PresentationQualifications And Experience Presentation
Qualifications And Experience Presentation
Kevin Baker
 
Perspective of pharmaceutical molecular design
Perspective of pharmaceutical molecular designPerspective of pharmaceutical molecular design
Perspective of pharmaceutical molecular design
Peter Kenny
 

Ähnlich wie Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile (20)

f33-ft-computing-lec02-measures.ppt
f33-ft-computing-lec02-measures.pptf33-ft-computing-lec02-measures.ppt
f33-ft-computing-lec02-measures.ppt
 
Qualifications And Experience Presentation
Qualifications And Experience PresentationQualifications And Experience Presentation
Qualifications And Experience Presentation
 
Molecular design: How to and how not to?
Molecular design:  How to and how not to?Molecular design:  How to and how not to?
Molecular design: How to and how not to?
 
Perspective of pharmaceutical molecular design
Perspective of pharmaceutical molecular designPerspective of pharmaceutical molecular design
Perspective of pharmaceutical molecular design
 
dfma_seminar
dfma_seminardfma_seminar
dfma_seminar
 
Computational Chemistry Robots
Computational Chemistry RobotsComputational Chemistry Robots
Computational Chemistry Robots
 
Optimization of Corridor Observation Method to Solve Environmental and Econom...
Optimization of Corridor Observation Method to Solve Environmental and Econom...Optimization of Corridor Observation Method to Solve Environmental and Econom...
Optimization of Corridor Observation Method to Solve Environmental and Econom...
 
NNPDF3.1
NNPDF3.1NNPDF3.1
NNPDF3.1
 
Molecular design: One step back and two paths forward
Molecular design:  One step back and two paths forwardMolecular design:  One step back and two paths forward
Molecular design: One step back and two paths forward
 
Real Time Geodemographics
Real Time GeodemographicsReal Time Geodemographics
Real Time Geodemographics
 
An Algorithm For Vector Quantizer Design
An Algorithm For Vector Quantizer DesignAn Algorithm For Vector Quantizer Design
An Algorithm For Vector Quantizer Design
 
Accelerated life testing
Accelerated life testingAccelerated life testing
Accelerated life testing
 
new optimization algorithm for topology optimization
new optimization algorithm for topology optimizationnew optimization algorithm for topology optimization
new optimization algorithm for topology optimization
 
USE OF ADAPTIVE COLOURED PETRI NETWORK IN SUPPORT OF DECISIONMAKING
USE OF ADAPTIVE COLOURED PETRI NETWORK IN SUPPORT OF DECISIONMAKINGUSE OF ADAPTIVE COLOURED PETRI NETWORK IN SUPPORT OF DECISIONMAKING
USE OF ADAPTIVE COLOURED PETRI NETWORK IN SUPPORT OF DECISIONMAKING
 
Phonons & Phonopy: Pro Tips (2014)
Phonons & Phonopy: Pro Tips (2014)Phonons & Phonopy: Pro Tips (2014)
Phonons & Phonopy: Pro Tips (2014)
 
BrazMedChem2014
BrazMedChem2014BrazMedChem2014
BrazMedChem2014
 
Universal approximators for Direct Policy Search in multi-purpose water reser...
Universal approximators for Direct Policy Search in multi-purpose water reser...Universal approximators for Direct Policy Search in multi-purpose water reser...
Universal approximators for Direct Policy Search in multi-purpose water reser...
 
Aeolian vibrations of overhead transmission line bundled conductors during in...
Aeolian vibrations of overhead transmission line bundled conductors during in...Aeolian vibrations of overhead transmission line bundled conductors during in...
Aeolian vibrations of overhead transmission line bundled conductors during in...
 
Some new directions for pharmaceutical molecular design
Some new directions for pharmaceutical molecular designSome new directions for pharmaceutical molecular design
Some new directions for pharmaceutical molecular design
 
Electric distribution network reconfiguration for power loss reduction based ...
Electric distribution network reconfiguration for power loss reduction based ...Electric distribution network reconfiguration for power loss reduction based ...
Electric distribution network reconfiguration for power loss reduction based ...
 

Mehr von Kathrin Dentler

Redundant Elements in SNOMED CT Concept Definitions
Redundant Elements in SNOMED CT Concept DefinitionsRedundant Elements in SNOMED CT Concept Definitions
Redundant Elements in SNOMED CT Concept Definitions
Kathrin Dentler
 
Semantic Integration of Patient Data and Quality Indicators based on openEHR ...
Semantic Integration of Patient Data and Quality Indicators based on openEHR ...Semantic Integration of Patient Data and Quality Indicators based on openEHR ...
Semantic Integration of Patient Data and Quality Indicators based on openEHR ...
Kathrin Dentler
 
Towards the Automated Calculation of Clinical Quality Indicators
Towards the Automated Calculation of Clinical Quality IndicatorsTowards the Automated Calculation of Clinical Quality Indicators
Towards the Automated Calculation of Clinical Quality Indicators
Kathrin Dentler
 
Approximation and Self-Organisation on the Web of Data
Approximation and Self-Organisation on the Web of DataApproximation and Self-Organisation on the Web of Data
Approximation and Self-Organisation on the Web of Data
Kathrin Dentler
 

Mehr von Kathrin Dentler (7)

Formalization and automated computation of diabetes quality indicators with C...
Formalization and automated computation of diabetes quality indicators with C...Formalization and automated computation of diabetes quality indicators with C...
Formalization and automated computation of diabetes quality indicators with C...
 
Computing Healthcare Quality Indicators Automatically: Secondary Use of Pati...
Computing Healthcare Quality Indicators Automatically:  Secondary Use of Pati...Computing Healthcare Quality Indicators Automatically:  Secondary Use of Pati...
Computing Healthcare Quality Indicators Automatically: Secondary Use of Pati...
 
Redundant Elements in SNOMED CT Concept Definitions
Redundant Elements in SNOMED CT Concept DefinitionsRedundant Elements in SNOMED CT Concept Definitions
Redundant Elements in SNOMED CT Concept Definitions
 
Semantic Integration of Patient Data and Quality Indicators based on openEHR ...
Semantic Integration of Patient Data and Quality Indicators based on openEHR ...Semantic Integration of Patient Data and Quality Indicators based on openEHR ...
Semantic Integration of Patient Data and Quality Indicators based on openEHR ...
 
The Reproducibility of CLIF, a Method for Clinical Quality Indicator Formali...
The Reproducibility of CLIF, a Method for Clinical Quality Indicator Formali...The Reproducibility of CLIF, a Method for Clinical Quality Indicator Formali...
The Reproducibility of CLIF, a Method for Clinical Quality Indicator Formali...
 
Towards the Automated Calculation of Clinical Quality Indicators
Towards the Automated Calculation of Clinical Quality IndicatorsTowards the Automated Calculation of Clinical Quality Indicators
Towards the Automated Calculation of Clinical Quality Indicators
 
Approximation and Self-Organisation on the Web of Data
Approximation and Self-Organisation on the Web of DataApproximation and Self-Organisation on the Web of Data
Approximation and Self-Organisation on the Web of Data
 

Kürzlich hochgeladen

Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
ZurliaSoop
 

Kürzlich hochgeladen (20)

Graduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - EnglishGraduate Outcomes Presentation Slides - English
Graduate Outcomes Presentation Slides - English
 
On_Translating_a_Tamil_Poem_by_A_K_Ramanujan.pptx
On_Translating_a_Tamil_Poem_by_A_K_Ramanujan.pptxOn_Translating_a_Tamil_Poem_by_A_K_Ramanujan.pptx
On_Translating_a_Tamil_Poem_by_A_K_Ramanujan.pptx
 
Single or Multiple melodic lines structure
Single or Multiple melodic lines structureSingle or Multiple melodic lines structure
Single or Multiple melodic lines structure
 
HMCS Vancouver Pre-Deployment Brief - May 2024 (Web Version).pptx
HMCS Vancouver Pre-Deployment Brief - May 2024 (Web Version).pptxHMCS Vancouver Pre-Deployment Brief - May 2024 (Web Version).pptx
HMCS Vancouver Pre-Deployment Brief - May 2024 (Web Version).pptx
 
ICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptxICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptx
 
Food safety_Challenges food safety laboratories_.pdf
Food safety_Challenges food safety laboratories_.pdfFood safety_Challenges food safety laboratories_.pdf
Food safety_Challenges food safety laboratories_.pdf
 
Towards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptxTowards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptx
 
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptxExploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
Exploring_the_Narrative_Style_of_Amitav_Ghoshs_Gun_Island.pptx
 
FSB Advising Checklist - Orientation 2024
FSB Advising Checklist - Orientation 2024FSB Advising Checklist - Orientation 2024
FSB Advising Checklist - Orientation 2024
 
SOC 101 Demonstration of Learning Presentation
SOC 101 Demonstration of Learning PresentationSOC 101 Demonstration of Learning Presentation
SOC 101 Demonstration of Learning Presentation
 
Plant propagation: Sexual and Asexual propapagation.pptx
Plant propagation: Sexual and Asexual propapagation.pptxPlant propagation: Sexual and Asexual propapagation.pptx
Plant propagation: Sexual and Asexual propapagation.pptx
 
80 ĐỀ THI THỬ TUYỂN SINH TIẾNG ANH VÀO 10 SỞ GD – ĐT THÀNH PHỐ HỒ CHÍ MINH NĂ...
80 ĐỀ THI THỬ TUYỂN SINH TIẾNG ANH VÀO 10 SỞ GD – ĐT THÀNH PHỐ HỒ CHÍ MINH NĂ...80 ĐỀ THI THỬ TUYỂN SINH TIẾNG ANH VÀO 10 SỞ GD – ĐT THÀNH PHỐ HỒ CHÍ MINH NĂ...
80 ĐỀ THI THỬ TUYỂN SINH TIẾNG ANH VÀO 10 SỞ GD – ĐT THÀNH PHỐ HỒ CHÍ MINH NĂ...
 
COMMUNICATING NEGATIVE NEWS - APPROACHES .pptx
COMMUNICATING NEGATIVE NEWS - APPROACHES .pptxCOMMUNICATING NEGATIVE NEWS - APPROACHES .pptx
COMMUNICATING NEGATIVE NEWS - APPROACHES .pptx
 
NO1 Top Black Magic Specialist In Lahore Black magic In Pakistan Kala Ilam Ex...
NO1 Top Black Magic Specialist In Lahore Black magic In Pakistan Kala Ilam Ex...NO1 Top Black Magic Specialist In Lahore Black magic In Pakistan Kala Ilam Ex...
NO1 Top Black Magic Specialist In Lahore Black magic In Pakistan Kala Ilam Ex...
 
Python Notes for mca i year students osmania university.docx
Python Notes for mca i year students osmania university.docxPython Notes for mca i year students osmania university.docx
Python Notes for mca i year students osmania university.docx
 
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
 
How to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptxHow to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptx
 
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdfUGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
UGC NET Paper 1 Mathematical Reasoning & Aptitude.pdf
 
Interdisciplinary_Insights_Data_Collection_Methods.pptx
Interdisciplinary_Insights_Data_Collection_Methods.pptxInterdisciplinary_Insights_Data_Collection_Methods.pptx
Interdisciplinary_Insights_Data_Collection_Methods.pptx
 
How to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POSHow to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POS
 

Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile

  • 1. Introduction Dimensions Results Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile WAI Kathrin Dentler November 8th 2010 Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 1/26
  • 2. Introduction Dimensions Results Outline 1 Introduction 2 Dimensions 3 Results Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 2/26
  • 3. Introduction Dimensions Results Outline 1 Introduction 2 Dimensions 3 Results Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 3/26
  • 4. Introduction Dimensions Results PhD Project Supervisors Ronald Cornet Annette ten Teije Nicolette de Keizer Frank van Harmelen Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 4/26
  • 5. Introduction Dimensions Results PhD Project SEMI4 SEMantic Interoperability for Integration of clinical Information and Indicators Quality Indicators General Problem: growing number, labour intensive Approach: formalize indicators; integrate data sources; investigate which adjustments are needed to facilitate data-capture at the required quality levels → More in next WAI Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 5/26
  • 6. Introduction Dimensions Results SNOMED CT Systematized Nomenclature of Medicine, Clinical Terms SNOMED CT used to represent information in electronic health records around 300,000 active concepts available as DB tables and as OWL ontology in tractable OWL 2 profile EL Goal: Classify SNOMED CT with a reasoner to obtain the closure, i.e. all possible inferences. Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 6/26
  • 7. Introduction Dimensions Results An example OWL 2 EL ontology Fracture Traumatic abnormality by morphology Traumatic abnormality by morphology Traumatic abnormality Traumatic abnormality Damage Fracture of foot ≡ Fracture of lower limb ∃.rolegroup (∃associated morphology.Fracture ∃finding site.Bone structure of foot) Fracture of lower limb ≡ Fracture of bone ∃rolegroup. (∃associated morphology.Fracture ∃finding site.Bone structure of lower limb) Inferred: Fracture Damage Fracture of foot Fracture of lower limb Pellet’s explanation: Fracture of foot ≡ Fracture of bone ∃.rolegroup (∃associated morphology.Fracture ∃finding site.Bone structure of foot) Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 7/26
  • 8. Introduction Dimensions Results Included Reasoners CB CEL FaCT++ HermiT Pellet Racer Snorocket TrOWL Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 8/26
  • 9. Introduction Dimensions Results Related Work Earlier Comparisons of Classification Performance for SNOMED CT 10000 CB CEL FaCT Racer Snorocket Classification Time in Seconds (log-scale) 1000 100 10 2005 2006 2008 2009 2010 Timeline Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 9/26
  • 10. Introduction Dimensions Results Outline 1 Introduction 2 Dimensions 3 Results Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 10/26
  • 11. Introduction Dimensions Results Dimensions 1 Reasoning Characteristics 2 Practical Usability 3 Performance Indicators Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 11/26
  • 12. Introduction Dimensions Results Dimension Reasoning Characteristics Independent of Ontology Methodology Soundness and Completeness in Theory Expressivity and Computational Complexity Incremental Classification Rule Support Justifications Support of ABox Reasoning Tasks Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 12/26
  • 13. Introduction Dimensions Results Dimension Practical Usability Independent of Ontology OWL API OWLlink Availability as Prot´g´ Plugin e e License Open or Closed Source Programming Language Supported Platforms Jena Kind of Institution Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 13/26
  • 14. Introduction Dimensions Results Dimension Performance Indicators Related to Ontology Classification Performance TBox Consistency Checking Performance Concept Satisfiability Checking Performance Subsumption Query Performance Required Heap Space Soundness and Completeness in Practice Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 14/26
  • 15. Introduction Dimensions Results Outline 1 Introduction 2 Dimensions 3 Results Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 15/26
  • 16. Introduction Dimensions Results Results: Categorization of Reasoners Dimension Reasoning Characteristics CB CEL FaCT++ HermiT Pellet RP SR TrOWL (REL) Methodology consequ.- compl. tableau- hyper tableau tableau compl. approx. based rules based tableau based based rules (compl. rules) Sound ( ) Complete x( ) IC (addition/ x/x /x x/x x/x / x/x /x x/x removal) Rule x x x x x Support (SWRL) (SWRL) (SWRL, nRQL) Justifications x x x x x ABox x x Reasoning (SPARQL) (SPARQL, (SPARQL) nRQL) Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 16/26
  • 17. Introduction Dimensions Results Results: Categorization of Reasoners Dimension Practical Usability CB CEL FaCT++ HermiT Pellet RP SR TrOWL OWL API x OWLlink API x x x Prot´g´ Plugin e e x x License DuLi: GLGPL AP 2.0 GLGPL GLGPL DuLi: AGPL own own DuLi: AGPL Open Source x x x Language OCaml Common Lisp C++ Java Java Lisp Java Java Platforms all Linux all all all all all all Jena x x x x x x x Institution a a a a c c g a Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 17/26
  • 18. Introduction Dimensions Results Results: Categorization of Reasoners Dimension Performance Indicators Experiments on Galjas With: (GO, NCI and) SNOMED CT Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 18/26
  • 19. Introduction Dimensions Results Classification Performance vs. Supported Expressivity Dimension Performance Indicators Supported Expressivity EL EL EL EL Horn SHIQ SHIQ(D-) SROIQ(D) SROIQ(D) G0 NCI SNOMED CT 1000 Classification Time in Seconds (log-scale) 100 10 1 0.1 CEL Pellet Snorocket TR CB RacerPro HermiT FaCT++ Reasoners ordered by increasing Supported Expressivity Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 19/26
  • 20. Introduction Dimensions Results Consistency Checking Performance Dimension Performance Indicators CB CEL FaCT++ HermiT Pellet RP SR TR OSNOMED CT x 0.88 15.3 0.00 16.78 x 0.00 0.00 CEL and TrOWL REL need to classify first, but return true anyway, Snorocket outputs a warning. Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 20/26
  • 21. Introduction Dimensions Results Concept Satisfiability Checking Performance Dimension Performance Indicators CB CEL FaCT++ HermiT Pellet RP SR TR OS CT BC x 38.42 22.37 5,276.85 56.91 273.45 0.07 5.17 OS CT AC x 34.59 1.76 1.36 6.07 0.00 0.06 0.46 E.g TrOWL returns reliable results only after the ontology is classified. → benchmarks with inconsistent ontologies / unsatisfiable concepts needed! Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 21/26
  • 22. Introduction Dimensions Results Subsumption Query Performance Dimension Performance Indicators CB CEL FaCT++ HermiT Pellet RP SR TR NC BC # results x 1 20 20 20 20 x 0 seconds x 0.96 701.79 6,649.85 2,793.31 3,380.67 NPE 0.17 AC BC # results x 1 1 20 20 20 x 0 seconds x 0.00 0.06 16.94 0.49 0.74 NPE 0.00 NC AC # results x 20 20 20 20 20 20 20 seconds x 0.00 0.00 0.00 0.00 0.70 0.00 0.28 AC AC # results x 1 1 20 20 20 20 0 seconds x 0.00 0.06 17.12 0.00 0.92 6.97 0.00 Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 22/26
  • 23. Introduction Dimensions Results Minimum Heap Space for Java Reasoners Dimension Performance Indicators CB CEL FaCT++ HermiT Pellet RP SR TR n/a n/a n/a 4.5 10 n/a 2.5 4 Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 23/26
  • 24. Introduction Dimensions Results Soundness and Completeness in Practice Dimension Performance Indicators CB: missing SubClassOf owl:Thing for top-level classes CEL: additional EquivalentClasses and EquivalentProperties for each class and property RacerPro: additional sub class owl:Nothing for all leaf concepts Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 24/26
  • 25. Introduction Dimensions Results Soundness and Completeness in Practice Dimension Performance Indicators Table: SNOMED CT: Missing / Additional inferred SubClassOf statements in regard to the Relationships Table CB CEL FaCT++ HermiT Pellet RP SR TR Missing 0 0 0 0 386 0 86 0 Additional 0 0 0 0 546 0 34 0 Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 25/26
  • 26. Introduction Dimensions Results Questions ? Kathrin Dentler k.dentler@amc.uva.nl k.dentler@few.vu.nl http://reasonerbench.net/ Kathrin Dentler — Comparison of Reasoners for large Ontologies in the OWL 2 EL Profile 26/26