ScienceGuardians

ScienceGuardians

Did You Know?

ScienceGuardians hosts editors too

Explainable tree-based machine learning modeling and optimization for intelligent oil feedstock formulation towards enhanced biodiesel production

Authors: Osamudiamhen Oiwoh,Andrew Nosakhare Amenaghawon,Stanley Aimhanesi Eshiemogie,Obinna Chuks Muonanu,Mathias Ikhenna Eliake,Chiagoziem Godswill Ndukwe,Oghenerukevwe Jeffrey Oghenehwosa,Ibhadebhunuele Gabriel Okoduwa,Osarieme Osazuwa,Jean Mulopo,Heri Septya Kusuma
Journal: Biomass and Bioenergy
Publisher: Elsevier BV
Publish date: 2026-2
ISSN: 0961-9534 DOI: 10.1016/j.biombioe.2025.108491
View on Publisher's Website
Up
0
Down
::

Of critical concern is a fundamental methodological flaw in your synergy analysis that compromises the core premise of your study. Your calculation of the synergistic Index (SI) in Equation (3) requires the biodiesel yields from the individual, pure oils (WVO, NSO, CSO) to serve as the baseline for comparison. However, your paper provides no experimental data for the biodiesel yield of any of these individual feedstocks under the transesterification conditions used for the blends.

How can you conclusively state that “SI values were greater than 1” and claim a “positive synergistic effect” without the essential control data from the pure oils? The entire statistical validation via bootstrap resampling is built upon these undefined SI values, rendering your central claim of synergy scientifically unsubstantiated. Without this evidence, on what basis can you assert that blending is beneficial, and does this not fundamentally undermine the justification for your optimally formulated ternary blend and the subsequent interpretations drawn from your SHAP analysis?
 
 

  • You must be logged in to reply to this topic.