Modelling the sensitivity of electrochemical sensors using minimal impedance features
Author(s)
Rajith, Shaheen
Zhang, Xin
Ghoreishizadeh, Sara
Type
Conference Paper
Abstract
We present a regression-based framework for lightweight quantitative estimation of nominal sensitivity in electrochemical sensors with gold electrodes. Features extracted from electrochemical impedance spectroscopy (EIS) are used to train and compare regression models of varying complexity, namely linear Huber regressor, and two tree-based ensemble models: XGBoost and Random Forest. The models are trained and tested on data collected from gold screen printed electrodes (SPE) and gold plated printed circuit board (PCB) electrodes. Results show that although ensemble models achieve better peak predictive performance (median R2 score up to 0.929 for SPE and 0.954 for PCB), comparable results can be achieved using the linear Huber model (median R2 score of 0.873 for SPE and 0.793 for PCB) with only three to five parameters compared to hundreds for the ensemble models. The closed form expression of the Huber model also enables straightforward hardware implementation, making this approach suitable for sensitivity estimation in resource constrained environments.
Date Acceptance
2026-05-01
Publisher
IEEE
Copyright Statement
Subject to copyright. This paper is embargoed until publication. Once published the author’s accepted manuscript will be made available under a CC-BY License in accordance with Imperial’s Research Publications Open Access policy (www.imperial.ac.uk/oa-policy).
Source
IEEE Biosensors 2026
Publication Status
Accepted
Start Date
2026-08-30
Finish Date
2026-08-04
Coverage Spatial
Bangkok, Thailand