A nonlinear feature extraction-based approach using manifold learning algorithms is developed in order to improve the classification accuracy in an electronic tongue sensor array. The developed signal processing methodology is composed of four stages: data unfolding, scaling, feature extraction, and classification. This study aims to compare seven manifold learning algorithms: Isomap, Laplacian Eigenmaps, Locally Linear Embedding (LLE), modified LLE, Hessian LLE, Local Tangent Space Alignment (LTSA), and t-Distributed Stochastic Neighbor Embedding (t-SNE) to find the best classification accuracy in a multifrequency large-amplitude pulse voltammetry electronic tongue. A sensitivity study of the parameters of each manifold learning algorithm is also included. A data set of seven different aqueous matrices is used to validate the proposed data processing methodology. A leave-one-out cross validation was employed in 63 samples. The best accuracy (96.83%) was obtained when the methodology uses Mean-Centered Group Scaling (MCGS) for data normalization, the t-SNE algorithm for feature extraction, and k-nearest neighbors (kNN) as classifier.

Electronic tongue-type sensor arrays are devices used to determine the quality of substances and seek to imitate the main components of the human sense of taste. For this purpose, an electronic tongue-based system makes use of sensors, data acquisition systems, and a pattern recognition system. Particularly, in the latter, machine learning techniques are useful in data analysis and have been used to solve classification and regression problems. However, one of the problems in the use of this kind of device is associated with the development of reliable pattern recognition algorithms and robust data analysis. In this sense, this work introduces a taste recognition methodology, which is composed of several steps including unfolding data, data normalization, principal component analysis for compressing the data, and classification through different machine learning models. The proposed methodology is tested using data from an electronic tongue with 13 different liquid substances; this electronic tongue uses multifrequency large amplitude pulse signal voltammetry. Results show that the methodology is able to perform the classification accurately and the best results are obtained when it includes the use of K-nearest neighbor machine in terms of accuracy compared with other kinds of machine learning approaches. Besides, the comparison to evaluate the methodology is made with different classification performance measures that show the behavior of the process in a single number.

Damage classification is an important topic in the development of structural health monitoring systems. When applied to wind-turbine foundations, it provides information about the state of the structure, helps in maintenance, and prevents catastrophic failures. A data-driven pattern-recognition methodology for structural damage classification was developed in this study. The proposed methodology involves several stages: (1) data acquisition, (2) data arrangement, (3) data normalization through the mean-centered unitary group-scaling method, (4) linear feature extraction, (5) classification using the extreme gradient boosting machine learning classifier, and (6) validation applying a 5-fold cross-validation technique. The linear feature extraction capabilities of principal component analysis are employed; the original data of 58,008 features is reduced to only 21 features. The methodology is validated with an experimental test performed in a small-scale wind-turbine foundation structure that simulates the perturbation effects caused by wind and marine waves by applying an unknown white noise signal excitation to the structure. A vibration-response methodology is selected for collecting accelerometer data from both the healthy structure and the structure subjected to four different damage scenarios. The datasets are satisfactorily classified, with performance measures over 99.9% after using the proposed damage classification methodology.