Background: The biosensors are analytical devices combining a bioreceptor and a physicochemical
transducer to translate the signal resulting from the interaction of the analyte with the biological
element into an electrical reading. The technological attractiveness of the biosensors resides in their high
sensitivity, continuous measuring capabilities, low cost, and precision. An important and promising area
of biosensors application is the associated with blood glucose levels monitoring in diabetic patients.
However, accuracy and stability issues are still under development, impeding the more widespread offering
of glucose biosensors in the market. Thus, this work aims to contribute to the biosensors technology
by analyzing the amperometric response to glucose of a typical glucose oxidase-based sensor of second
generation, applying mathematical models.
Methods: The amperometric response of the glucose oxidase-based electrochemical sensor of second
generation was analyzed under different conditions such as temperature, pH, mediators’ and substrate’s
concentrations through statistical learning methods from a regression perspective. Artificial Neural Networks
(ANN), LASSO and RIDGE Regression and Classification and Regression (CART) were the algorithms
selected to achieve the regression task.
Results: The reported experimental results show a promising very low prediction error of the biosensor
output by using Neural Networks and Classification and Regression Trees. It is also shown that the relationship
between predictors, i.e. features or variables- and response-target variable- corresponds to a nonlinear
Conclusion: A final CART and a Neural Network models are presented as a simple solutions that can be
solidly constructed in order to predict the amperometric response of a glucose oxidase sensor. The experimental
proposal and conditions offered in this paper could be applied for other scenarios in the wide
spectrum of biosensing technology.