Ion-Selective Electrode Array Based on a Bayesian Nonlinear Source Separation Method

Ion-selective electrodes (ISE) offer a practical approach for estimating ionic activities. Nonetheless, such devices are not selective, i.e., the ISE response can be affected by interfering ions other than the target one. With the aim of overcoming this problem, we propose a Bayesian nonlinear source separation method for processing the data acquired by an ISE array. The Bayesian framework permits us to easily incorporate prior information such as the non-negativity of the sources into the separation method. The effectiveness of our proposal is attested by experiments using artificial and real data.