A dual-target imprinted sensor integrating adaptive compensation and Monte Carlo-PCA-SVM intelligent algorithm for simultaneous detection of Hg 2+ and Pb 2+ in random water environments

A dual-target imprinted electrochemical sensor integrating adaptive differential compensation with a Monte Carlo-principal component analysis-support vector machine (Monte Carlo-PCA-SVM) intelligent algorithm is proposed for simultaneous, high-precision detection of Hg2+ and Pb2+ in randomly fluctuating water environments. The sensor adopts a dual-working-electrode architecture: an imprinted chitosan electrode yields target responses superimposed with environmental noise, whereas a template-free chitosan electrode responds only to common-mode noise; differential subtraction achieves adaptive noise cancellation without dedicated temperature or pH sensing channels. Owing to limited labelled data from disposable imprinted sensors, Monte Carlo simulation augments the training set, PCA reduces high-dimensional features via orthogonal projection, and an SVM classifier performs dual-target recognition, with kinetic analysis further elucidating the reaction mechanism. Under temperature fluctuation, adaptive compensation reduces detection errors of Pb2+ and Hg2+ from 26.41% and 23.56% to 6.38% and 10.47%, respectively, and recognition accuracy reaches 96.91% (σ = 0.5). The sensor exhibits a linear range of 20–100 nM with detection limits of 5.25 nM (Pb2+) and 7.28 nM (Hg2+), and recoveries of 96.73%–100.16% in tap water. This work provides an auxiliary-sensor-free compensation strategy and a data-driven recognition framework for multi-target heavy-metal monitoring in dynamic aquatic environments.