Locally Linear Embedding as Nonlinear Feature Extraction to Discriminate Liquids with a Cyclic Voltammetric Electronic Tongue

Electronic tongues are devices used in the analysis of aqueous matrices for classification or quantification tasks. These systems composed several sensors different materials, a data acquisition unit, and pattern recognition system. Voltammetric have been electronic using cyclic voltammetry method. By this method, each sensor yields voltammogram that relates response current to change voltage applied working electrode. A great amount is obtained experimental procedure which allows handling as application; however, development efficient machine-learning-based methodologies still an open research interest topic. As contribution, work presents novel processing methodology classify signals acquired by voltammetric tongue. This stages such normalization through group scaling method nonlinear feature extraction step with locally linear embedding (LLE) technique. The reduced-size vector input k-Nearest Neighbors (k-NN) supervised classifier algorithm. leave-one-out cross-validation (LOOCV) performed obtain final accuracy. validated set five juices liquid substances.Two screen-printed electrodes voltametric were Specifically materials their platinum graphite. results reached 80% accuracy after applying developed methodology.