A Bayesian Conditional Diffusion Regression Framework with Hierarchical Global-Local Priors for Robust Agricultural Image-to-Yield Prediction

Kumar Ch; Upender reddy G; Varalakshmi K

doi:10.5269/bspm.82344

A Bayesian Conditional Diffusion Regression Framework with Hierarchical Global-Local Priors for Robust Agricultural Image-to-Yield Prediction

Authors

Kumar Ch
Upender reddy G Univeristy college of Engineering(A), Osmania Univeristy,
Varalakshmi K MJPTBCWR Degree College (W), Wargal, Siddipet-502279, Telangana, India.

DOI:

https://doi.org/10.5269/bspm.82344

Abstract

Accurateestimationofcropyieldfromimage-deriveddataremainsdifficultduetothecombined
effectsofhighpredictordimensionality,stronginter-featuredependence,non-standardnoisebehavior,andthe
presenceofanomalousobservations.Manycommonlyusedregressionapproaches, includingBayesianmodels
withGaussianerrorassumptions,struggletoremainstableundertheseconditionsandoftenprovideunreliable
uncertaintyestimates.ThisworkpresentsBayesianConditionalDiffusionRegression(BCDR),aprobabilistic
modelingframeworkthatapproachesregressionthroughconditional responsegeneration. Themethodinte
gratesadiffusion-basedmechanismformodelingtheresponsevariablewithhierarchicalglobal–localshrinkage
tocontrolmodelcomplexity,whileaheavytailedlikelihoodimprovesrobustnesstooutliers. Posterior infer
enceiscarriedoutusingahybridGibbsandHamiltonianMonteCarlostrategytoenabletractableestimation
inhigh-dimensionalsettings.Experimentsconductedonsyntheticdatasetsandarealagriculturalcasestudy
involvingpomegranateimagesandfruitweightmeasurementsdemonstrateimprovedpredictiveaccuracy,bet
teruncertaintycalibration, andstronger resistancetodatacontaminationwhencomparedwithestablished
statisticalandmachine-learningbaselines.TheseresultssuggestthatBCDRisarobustBayesianframework
forimage-basedyieldprediction

Downloads

Published

2026-06-19

Issue

Vol. 44 No. 17 (2026): Recent Trends in Mathematical Sciences and Technological Applications

Section

Conf. Issue: Recent Trends in Mathematical Sciences and Technological Applic.

License

This work is licensed under a Creative Commons Attribution 4.0 International License.

When the manuscript is accepted for publication, the authors agree automatically to transfer the copyright to the (SPM).

The journal utilize the Creative Common Attribution (CC-BY 4.0).

How to Cite

Ch, K., G, U. reddy, & K, V. (2026). A Bayesian Conditional Diffusion Regression Framework with Hierarchical Global-Local Priors for Robust Agricultural Image-to-Yield Prediction. Boletim Da Sociedade Paranaense De Matemática, 44(17), 1-13. https://doi.org/10.5269/bspm.82344