Dataset for the article: 'Modeling Women Cyclists' Perceived Security: A Comparison of Machine Learning Techniques'

This dataset contains data collected from female participants assessing perceived security while cycling in a virtual reality (VR) laboratory at the University of Tehran, Iran, as described in the study "Modeling Women Cyclists' Perceived Security: A Comparison of Machine Learning Techniques." Participants experienced multiple VR scenarios simulating various urban environments and traffic conditions, with features including crowdedness (1-4), incivility (1-3), lighting (1-3), obstruction of visibility (1-4), and surveillance (0-1). After each scenario, participants rated their perceived security (1=low, 2=medium, 3=high) via a questionnaire. The final dataset includes 208 samples after outlier removal, with five independent variables and the target variable "perceived security."
Data was preprocessed with normalization and standardization for machine learning analysis. See the accompanying README.txt and the paper for variable definitions and file structure. This dataset supports the development of the DNN-stack2 model (84.07% accuracy) and is intended for research on cyclist security, urban planning, and machine learning applications.