Athlete-Independent Capoeira Movement Recognition: Data and Reproducibility Materials

This dataset supports the manuscript “Athlete-Independent Capoeira Movement Recognition Using Markerless Human Pose Estimation and Machine Learning”. The repository contains de-identified derived data and reproducibility materials, including the archived 64-step temporal matrix, corrected Random Forest feature matrix, athlete partitions and development folds, feature definitions, sequence-level development and held-out predictions, reconstruction and evaluation scripts, and numerical verification reports. Identifiable audiovisual recordings and personal participant information are not included because of privacy and consent restrictions.