logseq/pages/Meetings/Mentoring/2021_01_16-1024-Meeting-Riccardo-Juri.md

2021_01_16-1024-Meeting-Riccardo-Juri

#meeting

---

202101161024-Meeting-Riccardo-Juri

Date: 2021-01-16 Time: 10:24 Attendend: #attendance/online Project: #recsys #modeling People attended: #people/juri #people/riccardo

---

Overview dell'approccio come presentato da Riccardo

!

Molto simile al paper AMMORE.pdf. La differenza è che la query viene generata automaticamente.

Overview dell'approccio proposto

graph TB


 input(Metamodel being developed)
 tool[Lucene-based tool]
 output(OCL Statements)
 rankingTool[Ranking tool]
 rankingPolicy(Ranking policy)
 rankedStatements(Ranked OCL Statements)


 input-->tool
 tool-->output
 output-->rankingTool
 rankingPolicy-->rankingTool
 rankingTool-->rankedStatements

User input: It consists of the current metamodel being developed, which includes OCL statements

Input of the approach: It consists only of the metamodel being specified.

Lucene-based tool: It retrieves: 1) metamodels that similar to the one given in input 2) The metamodels retrieved in 1) are used to retrieve OCL statements that can be used to complete the specification of the metamodel being developed.

Ranking tool: It takes as input the retrieved OCL statements, a ranking policy, and it generates the ranked list of OCL statements.

Evaluation

For the Evaluation we apply the following process:

1. We create a dataset of metamodels containing OCL statements (this represents the GT)
2. We remove from the dataset in 1) all the OCL statements
3. We apply the approach on all the metamodels in 2) and compare the recommended OCL statements with those in the GT.
   1. Issue: we need to identify the similarity function to be used for comparison.