Soft computing techniques in modeling the influence of pH on dopamine biosensor

Recently the soft computing techniques have become an important alternative tool to conventional methods in modeling complex non-linear relationships. Since the electrochemical biosensors use an enzyme reaction for measuring different types of substrates, the pH and temperature influence strongly on the output signal of biosensors. The paper treats soft computing modeling the input/output non-linear dependence of a dopamine biosensor, which uses an active membrane from banana tissue. The model represents the biosensorpsilas output current versus the substrate concentration and pH. The temperature is not taken into account, it is set to be constant during the experiments. The problem to solve here is to find a way of increasing the accuracy (and the fastness) of the modeling process, under condition of insufficient experimental data. The following soft computing techniques are compared in MATLAB environment: (1) neural network with back-propagation learning algorithm, (2) CMAC neural network, (3) fuzzy logic, and (4) ANFIS. The relative errors over a few new experimental samples are calculated for validation of the proposed models.