4 Alin is a system for interactive ontology matching. In last year's version, Alin applied natural language processing techniques (NLP) to standardize the concept names of the ontologies that participate in the matching process. In the current version, we modi ed the grammars used during the process in order to improve the quality of the alignments. This article describes the participation of Alin at OAEI 2021 and discusses its results.
Alignment Evaluation Initiative). Alin [
Interactive ontology matching systems select mappings for domain expert evaluates. Alin selects many of these mappings through semantic and lexical metrics. As the concept names of the ontologies are not standardized, these metrics may return lower values than would be the case if they were standardized. This smaller metric may cause Alin not to select these mappings for evaluation by the domain expert.
In its 2020 version, Alin proposed the use of Natural Language Processing (NLP) resources such as regular grammars (in fact, their equivalent regular expressions) and context-free grammars along with their respective lexical analyzers (scanners) and syntax analyzers (parsers).
The use of these NLP resources makes it possible to translate di erent patterns used in the two ontologies into a unique one. This standardization allows Alin to select better mappings for the domain expert to evaluate.
To do the standardization, Alin has a new phase before the program runs. In this phase, an NLP expert manually develops grammars, and their respective scanners and parsers, to the ontologies. Alin uses these scanners and parsers during the execution of the program. This new phase is possible in an interactive ontology matching system because: 1. We know before the program runs which ontologies it will match, as we need to look for experts in the domain of ontologies to interact with the program; 2. The process of searching, meeting, and scheduling a day available for the expert to participate in the process can take a long time, probably a few days.
We can use this time of a few days until the execution of the program to develop the necessary grammars, scanners, and parsers for the ontologies. To test the new techniques, the authors of this paper played the role of the NLP experts. 1.2
During its matching process, Alin handles three sets of mappings: (i) Accepted, which is a set of mappings de nitely to be retained in the alignment; (ii) Selected, which is a set of mappings where each is yet to be decided if it will be included in the alignment; and (iii) Suspended, which is a set of mappings that have been previously selected, but (temporarily or permanently) ltered out of the selected mappings.
Given the previous de nitions, Alin procedure follows 5 Steps, described as follows: 1. Select mappings: select the rst mappings and automatically accepts some of them. We explain the selection and acceptance techniques in the 'Description of the used techniques' paragraph below; 2. Filter mappings: suspend some selected mappings, using lexical criteria for that; 3. Ask domain expert: accepts or rejects selected mappings, according to domain expert feedback 4. Propagate: select new mappings, reject some selected mappings or unsuspend some suspended mappings (depending on newly accepted mappings) 5. Go to step 3 as long as there are undecided selected mappings
All versions of Alin (since its very rst OAEI participation) follow this
general procedure. In this 2020 version, Alin included a new step where an
NLP expert develops grammars, and their respective scanners and parsers to
the concept names of the ontologies. Alin uses these scanners and parsers to
standardize the concept names of the ontologies and thus improve the generated
alignment. The new step can lead to, for example, correcting spelling errors and
unifying di erent spellings for the same concept name. More detailed examples
of possible standardization of concept names are presented in [
{ Step 1. Alin runs the scanners and the parsers for each concept name of the
ontologies, modifying it and standardizing it. Alin uses a blocking strategy
where it discards all data properties and object properties of the ontologies.
So, in this step, Alin selects only concept mappings, using linguistic
similarities between the standardized concept names. Alin automatically accepts
concept mappings whose standardized names are synonyms. Alin uses the
Wordnet and domain-speci c ontologies (the FMA Ontology in the Anatomy
track) to nd synonyms between entities.
{ Step 2. Alin suspends the selected mappings whose entities have low lexical
similarity. We use the Jaccard, Jaro-Wrinkler, and n-gram lexical metrics
to calculate the lexical similarity of the selected mappings. We based the
process of choosing the similarity metrics used by Alin on the result of
these metrics in assessments [
{ Step 5. The interaction phase continues until there are no selected mappings. Mapping anti-patterns An anti-pattern mapping can be a logical inconsistency, a construction constraint on the ontology, or an alignment constraint. An ontology may have construction constraints, such as a concept cannot be equivalent to its superconcept. The alignment between two ontologies can have a constraint like, for example, an entity of ontology O cannot be equivalent to two entities of the ontology O0. Anti-pattern mapping is a combination of mappings that generates a problematic alignment, i.e., a logical inconsistency or a violated constraint.
In its 2021 version, we modi ed the grammars used by Alin to improve the quality of the generated alignments. 1.3
Alin is available 6 as a package to be run through the SEALS client. 2
Results Interactive ontology matching is the focus of the Alin system.
The comparison between the participation of Alin in 2021 and 2020 (Tables 4 and 5) shows an improvement in the quality of the generated alignment, attesting the e ectiveness of the changes in the grammars. 2.1
The modi ed grammars generated an increase in the F-Measure of non-interactively generated alignments in the Anatomy track. The modi ed grammars generated a minor increase in the F-Measure of non-interactively generated alignments in the Anatomy track. There was a decrease in precision and an increase in recall. The current version of Alin did not participate in the Conference track. (Table 1). 6 https : ==drive:google:com=f ile=d=1SxJ L6f LRV qI84epm8DbAM lcscEoGbgZ=view?usp = sharing
In the Anatomy track, Alin was better than LogMap in quality (F-Measure) but worse in total requests, and worse in both aspects than AML (Table 2). In the Conference track, Alin was rst in quality and third in total requests (Table 3). Alin improved the quality of the alignment generated for both the Anatomy and the Conference tracks, with a higher improvement for the Anatomy scenario. The number of interactions has also increased on both tracks (Tables 4 and 5).
The quality of the alignment generated by Alin depends on the correct feedback from the domain expert, as Alin uses this feedback to select new mappings. When Alin selects wrong mappings, the quality of the generated alignment tends to decrease. The comparison between the two runs of the table 6 shows a quality decline sharper in the current version.
The run time of Alin 2021 was shorter than last year's (Table 7), but other systems also had a similar reduction. So there is a chance that this reduction may be due to changes in the computational environment. The analysis of OAEI 2021 results showed that Alin 2021 has improved the quality of the generated alignment in the interactive track. Moreover, an increase in the expert error rate caused a sharper fall in the alignment quality. The number of interactions with the expert was increased (from 233 to 281 requests in the Conference track and from 360 to 404 requests in the Anatomy track). Finally, Alin had a decrease in run time, like all other participants (Table 7), so the drop seems to be caused by changes in the computing environment rather than changes made to the program. This paper presented the newest version of Alin (Alin 2021), an interactive matching tool that makes use of natural language processing (NLP) resources, and its performance results obtained in OAEI 2021. When compared to its previous version, Alin 2021 used modi ed grammars during the NLP process to allow for a higher level of standardization among the terms used by the ontologies being aligned. These grammar modi cations were tested on the Anatomy and Conference tracks and proved to be e cient in increasing the quality of the generated alignment, obtaining the highest F-measure value in the Conference interactive track, with an expert error rate of zero.
In future versions, we intend to implement new grammar changes to further improve the quality of the generated alignment.