Behavioral Classification Dataset Analysis Code for Human-AI Decision Systems
This repository provides the analytical dataset, feature dictionary, cohort summaries, model specifications, and executable Python analysis workflow supporting reproducible analysis of behavioral patterns and cognitive accountability failures in human-AI decision systems across controlled laboratory, flight simulation, and operational aviation settings. The repository contains the analytical data structure used for classification experiments, including 127 engineered behavioral features, outcome labels, cohort information, feature definitions, model specifications, and reproducibility documentation. The analysis workflow implements the reported machine-learning methodology, including data partitioning, class-imbalance handling, Random Forest, Gradient Boosting, Logistic Regression, ensemble prediction, stacking, early prediction, study-specific evaluation, grouped validation, and domain-shift evaluation. The repository is intended to provide a transparent computational record of the analytical workflow and facilitate independent examination of the reported methodology and model evaluation procedures. Operationally sensitive, confidential, or individually identifiable source records are not included.
Authors
- Ghanshyam Ghatole
Institutions
- Woxsen School of Business (IN)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.5281/zenodo.24335332
- Primary Topic
- Human-Automation Interaction and Safety
- Type
- article
- Field-Weighted Citation Impact
- 0.00