20140204-NOOR

Invalid experiment: incorrect reference alignment (see 20140208-NOOR)

Experiment design

Experimenter: Jérôme Euzenat (INRIA)

Date: 2014-02-04

Hash: 769936317dba7a3a3d7294155ef333e12f5d905a

Hypothesis:

This the same experiment as the 20140131-NOOR: 3 agents performing revision of a network of ontology.

Experimental setting:

The NOOEnvironment and AlignmentAdjustingAgent have been rewritten for:

• using the OWL API and HermiT (instead of Jena);
• computing the reference alignment;
• computing Syntactic F-measures (PRecEvaluation).

AverageLogger has been provided to aggregate several runs together and compute the average.

Class used: NOOEnvironment, AlignmentAdjustingAgent, AlignmentRevisionExperiment, ActionLogger, AverageLogger, Monitor.

Execution environment: MacBook Pro 15, C2D 2.66GHz, 8GB - MacOS X 6.8 - Java 1.6.0 HotSpot

Result exploration

The following plot provides results using such features:

• In blue the average success rate (over 5 runs);
• In dotted, the minimum and maximum success rate (over the same 5 runs);
• In red, the average syntactic F-measure.

Analysis

Initial F-measure is .12 (likely corresponding to the 6 correct top-top correspondences), final is .15 (likely corresponding to the increase in precision once discarded all bad correspondences).

Further experiments:

With respect to F-measure, there are several things to be changed:

• Using SemPRecEvaluator (this requires to create the alignments through Ontowrap);
• Using a revision modality different from simply deleting the incorrect correspondences (this is already implemented).

This file can be retrieved from URL https://sake.re/20140204-NOOR

It is possible to check out the repository by cloning https://felapton.inrialpes.fr/cakes/20140204-NOOR.git

This experiment has been transferred from its initial location at https://gforge.inria.fr (not available any more)