Automating the Synthesis of Recommender Systems for Modelling Languages
summary
Automates generation of recommender systems for modelling languages via a textual configuration DSL that emits a deployable service and editor integration. Droid DSL specifies meta‑model, item types, recommendation methods and evaluation; generator uses Xtext, JET templates and produces a Java REST service (Jersey/Tomcat) with EMF tree editor integration. Offline evaluation on UML models from MAR uses 10‑fold CV and varied neighbourhoods, finding user-based collaborative filtering (n=10, F1=0.060); also shows integration with an existing modelling chatbot.
AI-written summary · about
chapters (16)
- 0:00Automating the Synthesis of Recommender Systems for Modelling Languages
- 0:22Motivation
- 1:30Overview of the Approach
- 2:10The Droid DSL
- 5:08Architecture
- 6:04Tool support: The Droid configurator
- 6:51Tool support: The Droid service
- 7:34Tool support Integration with EMF tree editor
- 8:12Evaluation: Usefulness of recommendations • Experiment setup
- 9:01Results of the experiment - Part
- 10:48Case study on RS integration
- 11:08RS integration in SOCIO
- 12:11Answering RQ2
- 12:34Conclusions
- 12:59Future work
- 13:18Question?
related talks
23mReuse and Automated Integration of Recommenders for Modelling Languages
29mIntegrating Language Engineering into the Software Development Process
15mAutomated Variability Injection for Graphical Modelling Languages
13mTowards Supporting SPL Engineering in Low-Code Platforms using a DSL Approach
27mOperationalizing the Integration of User Interaction Specifications in Modeling Editor Synthesis
29mApplying Self-Extension Mechanism to DSLs for Establishing Model Libraries
38mDSLs, Code Generation, and New Domains for Ruby
05mModel Driven Software: Putting the Cart Before the Egg