Vol. 14 (10) pp. 62-76 DOI: 10.21474/IJAR01/24314

PREDICTING STRENGTH FROM DENSITY IN SELECTED AFRICAN TIMBERS: ACCURACY AND SENSITIVITY OF SIMPLE MODELS

  • Institut Polytechnique, Universite Gamal Abdel Nasser de Conakry (UGANC), BP 1147, Conakry, Guinea .
9 Downloads 251 Views
Crossref

Abstract

Methods: Thirteen reference means from Tropix (CIRAD, 2023-2026), spanning 380-1060 kg/m3, were analysed. Proportional, affine and power models were compared with a mean-only baseline using internal leave-one-out and leave-one-genus-out cross-validation. Twelve genera were represented, with only one repeated genus. Deterministic deletion scenarios and influence diagnostics assessed sensitivity. Results: The proportional model achieved internal genus-out RMSE values of 8.96 MPa in bending and 2.97 MPa in compression, equivalent to 9.47% and 5.27% of the respective mean responses. These are errors against held-out documentary means, not laboratory measurement uncertainties or specimen-level prediction errors. Relative to the mean-only baseline, RMSE decreased by 73.2% and 84.4%. Density remained informative across the examined deletion scenarios, although model-form rankings could reverse.

Keywords

How to Cite This Article

Djidoula Tako et,al (2026); PREDICTING STRENGTH FROM DENSITY IN SELECTED AFRICAN TIMBERS: ACCURACY AND SENSITIVITY OF SIMPLE MODELS, International Journal of Advanced Research (IJAR), 14 (10), 62-76, ISSN 2320-5407. DOI: https://doi.org/10.21474/IJAR01/24314

Corresponding Author

Djidoula Tako
Institut Polytechnique, Universite Gamal Abdel Nasser de Conakry (UGANC), BP 1147, Conakry, Guinea .
Togo

Publication Achievement

Publication card

Share your publication card

Post this card with your article title and abstract to your professional network.

🎉 Proud to contribute to scientific research! Share your publication with your colleagues and professional network.

Article Analytics