Studies Dataset for "Artificial Intelligence for Source Code Understanding Tasks: A Systematic Mapping Study"

This dataset contains the collection and classification of primary studies analyzed in the paper “Artificial Intelligence for Source Code Understanding Tasks: A Systematic Mapping Study.” It includes metadata of research papers (title, authors, year, venue), categorized by research tasks (e.g., detection, search, summarization, generation, understanding), as well as information on applied models, embeddings, preprocessing techniques, and evaluation metrics.

The dataset was created to provide transparency and reproducibility of the systematic mapping process, and to support future research on artificial intelligence methods for source code understanding. Researchers and practitioners may reuse this dataset for meta-analysis, replication studies, benchmarking, or as a starting point for new literature reviews.

Contents:

Paper metadata (title, authors, year, venue)

Task classification (detection, search, summarization, generation, understanding)

AI models and techniques used

Embedding and preprocessing approaches

Evaluation strategies and datasets