Supplementary data for "A Systematic Review of Fault Prediction in Information Systems Using Big Data Analytics and Large-Scale AI Models": screening decisions, data extraction, and search log

This dataset contains the complete supplementary material for a systematic literature review of fault prediction in information systems, covering publications from 2020 to 2026. The review examines two research trajectories that have developed largely independently: code-level software defect prediction and operational failure prediction, including AIOps and log-based anomaly detection. Three files are provided. S1 records the screening decision for every one of the 2,399 unique records retrieved by the search, together with the exclusion criterion applied and the rationale recorded at the time of the decision. S2 contains the extracted data for the 62 studies read in full, comprising fault taxonomy, data sources, model architecture, evaluation practice, deployment evidence, author-stated limitations, and quality assessment scores. S3 documents the search strategy in full, including both the final and the superseded query formulations, per-system and per-year result counts, the complete deduplication chain from 3,282 identified records to the 680 studies included in the review, the two search systems evaluated and rejected, and a coverage check against a sample of prominent publications. Each file includes a README sheet describing its columns and the methodological caveats attaching to them. The material is released to permit independent verification of the screening and extraction procedures reported in the associated article.