Data and code for: From kinetics to machine learning: predictive assessment of oxidative stability and shelf life in edible oils, a systematic review

Data and analysis code for a PRISMA 2020 systematic review of 247 studies (2012–2026) that estimate, infer or forecast the oxidation of edible oils and oil-rich foods using kinetic models, spectroscopy, sensors and machine learning. The archive contains: - data/included_studies_247.csv – extraction and appraisal record for all included studies, including the eight appraisal item scores (adapted from REFORMS), study scores and quality bands - data/coded_dataset_247.csv – derived synthesis codes (method family, prediction task, ML use, oxidation context, oil group, target indices) - data/screening_records_808.csv – title/abstract screening records and decisions - data/fulltext_excluded_21.csv and data/not_retrieved_5.csv – full-text exclusions with reasons and reports not retrieved - data/linked_reports_19.csv – linked report groups - data/citation_chaining_candidates_86.xlsx – backward citation chaining candidates - data/appraisal_rule_check_corrections_75.csv – item scores corrected during the rule check - data/Supplementary_Tables_S1-S5.xlsx – supplementary tables of the article - code/ – Python scripts that reproduce every count, statistical test and figure reported in the article (see README.md) Screening, extraction and appraisal were done by a single reviewer assisted by an AI tool (Claude, Anthropic), with all AI-proposed decisions checked by the author.