Towards automatic extraction of UML class diagrams: Creation of an annotated dataset for training deep models
Bibliographic record
Abstract
Software modeling relies heavily on UML class diagrams, essential tools for structuring a system’s entities, behaviors, and relationships. Yet, manually developing them from textual specifications remains a time-consuming task and subject to interpretation. This study proposes the creation of a corpus annotated according to a customized IOB schema, intended to train Named Entity Recognition (NER) models for the automatic extraction of UML elements from text. The schema integrates specific labels to accurately capture classes, attributes, methods, and relationships (association, aggregation, composition, inheritance), including their compound forms. The current corpus, built from 132 documents from various sources, includes more than 900 sentences and 11,000 manually annotated tokens. Particular attention was paid to the syntactic and semantic diversity of the texts, as well as to the linguistic quality, to ensure good generalization of the models. The empirical evaluation conducted with six Transformers models (BERT, RoBERTa, SpanBERT, XLNet, MiniLM and Electra) shows promising results, especially for classes and their relationships. This work thus lays the foundation for a reliable automation of UML class diagram generation from textual specifications, with strong potential for integration into software engineering environments and MDA processes.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.010 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.006 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.003 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.006 | 0.006 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".