Summary
A PRISMA 2020 systematic review and three-level meta-analysis of 1,102 effect sizes: random cross-validation inflates R² by 0.21.
Back to publications
Q1October 2026 · Submitted
Random Cross-Validation Overestimates the Accuracy of Machine-Learning Environmental Maps: A Meta-Analysis of 266 Studies
International Journal of Geographical Information Science · Taylor & Francis
Lab authors
Related publications
Accepted · in press
2026·SpringerL3
Ritmo: A Privacy-Preserving Smartphone App and Live Dashboard for Passive Behavioral Sensing on Mid-Range Android Devices
Dina Maribel Yana-Yucra, Renato Quispe-Vargas, Richar Andre Vilca-Solorzano
Accepted · in press
2026·SpringerL3
Survey-Weighted Longitudinal Analysis of Preschool Shared Reading and Adolescent Vocabulary in Chile
Renato Quispe-Vargas, Ana Madeley Yana-Yucra, Richar Andre Vilca-Solorzano, Dina Maribel Yana-Yucra