China
Africa
Explore chapters and articles related to this topic
Applications of Machine Learning in Industrial Sectors
Published in Pedro Larrañaga, David Atienza, Javier Diaz-Rozo, Alberto Ogbechie, Carlos Puerto-Santana, Concha Bielza, Industrial Applications of Machine Learning, 2019
Pedro Larrañaga, David Atienza, Javier Diaz-Rozo, Alberto Ogbechie, Carlos Puerto-Santana, Concha Bielza
More practical goals within the chemical industry are to help create new and novel fragrances that perform well technically (do not irritate the skin, do not change color after a few months), smell good and are unique in the marketplace. Master perfumers typically train for ten years before they become proficient. Goodwin et al. (2017) use different classifiers to predict target gender and average rating for unseen fragrances, all characterized by a set of fragrance notes. Also, the data are projected in a 2D space using t-distributed stochastic neighbor embedding (t-SNE) (van der Maaten and Hinton, 2008), a non-linear dimensionality reduction technique. This discovers clusters of perfumes and free spaces without data, which could suggest combinations of as yet unexplored but promising fragrance notes. Xu et al. (2007) design a new pigment mixing method based on an artificial neural network to emulate real-life pigment mixing, as well as to support the creation of new artificial pigments.
Dimensionality Reduction — Nonlinear Methods
Published in Wendy L. Martinez, Angel R. Martinez, Jeffrey L. Solka, Exploratory Data Analysis with MATLAB®, 2017
Wendy L. Martinez, Angel R. Martinez, Jeffrey L. Solka
This chapter covers various methods for nonlinear dimensionality reduction, where the nonlinear aspect refers to the mapping between the high-dimensional space and the low-dimensional space. We start off by discussing a method that has been around for many years called multidimensional scaling. We follow this with several more recently developed nonlinear dimensionality reduction techniques called locally linear embedding, isometric feature mapping, and Hessian eigenmaps. We conclude this chapter by discussing some methods from the machine learning and neural network communities, such as self-organizing maps, generative topographic maps, curvilinear component analysis, autoencoders, and stochastic neighbor embedding.
Dimensionality Reduction and Feature Extraction and Classification
Published in Arturo Román Messina, Data Fusion and Data Mining for Power System Monitoring, 2020
Nonlinear dimensionality reduction methods can be broadly characterized as selection-based and projection-based. Among these methods, spectral dimensionality reduction (manifold-based) techniques have been successfully applied to power system data. Other dimensionality reduction tools, such as Koopman-based techniques and Markov algorithms are rapidly gaining acceptance in the power system literature (Ramos and Kutz 2019; Messina 2015).
Joint L2,p-norm and random walk graph constrained PCA for single-cell RNA-seq data
Published in Computer Methods in Biomechanics and Biomedical Engineering, 2023
Tai-Ge Wang, Jun-Liang Shang, Jin-Xing Liu, Feng Li, Shasha Yuan, Juan Wang
The T - Distribution Stochastic Neighbour Embedding (t-SNE) (Van der Maaten and Hinton 2008) is a nonlinear dimensionality reduction method that captures the local structure information of high-dimensional data while revealing the global structure. Due to its intuitive and easy-to-interpret visualization, t-SNE is very popular in scRNA-seq data analysis and serves a vital role for scientists in identifying cell subgroups. We set σ2 = 0.2; and then apply the principal direction matrix obtained by RWPPCA, gLPCA, DGPCA, and PgLPCA to t-SNE to observe the separation of cell clusters with each of the different methods. Figure 3 shows the clustering visualization results of the four methods on simulated data, where each dot represents one cell, and each color represents a cell type.
NNNPE: non-neighbourhood and neighbourhood preserving embedding
Published in Connection Science, 2022
Kaizhi Chen, Chengpei Le, Shangping Zhong, Longkun Guo, Ge Xu
Nonlinear dimensionality reduction approaches, in contrast to linear dimensionality reduction techniques, deal with complex nonlinear data, thus attracting widespread attention. Many nonlinear dimensionality reduction algorithms have been proposed in recent decades, such as isomaps (Tenenbaum et al., 2000), LLE (Roweis & Saul, 2000), Laplacian eigenmaps (LE) (Belkin & Niyogi, 2001), Hessian LLE (Donoho & Grimes, 2003), and LTSA (Zhang & Zha, 2004). These algorithms utilise nonlinear low-dimensional manifolds from sample datasets that are inherent in high-dimensional space. Isomap is a global approach in low-dimensional space that seeks to retain pairwise geodesic distances among data points. By contrast, other techniques are local methods. LLE and LE endeavour to keep the local geometry of data, and neighbour points on the high-dimensional are regarded as neighbouring on the low-dimensional manifold. Hessian LLE is similar to LE in that it replaces the manifold Laplacian with the manifold Hessian. Meanwhile, LTSA is a technique that uses the local tangent space of each sample to characterise the local features of high-dimensional data (Van Der Maaten et al., 2009). These nonlinear dimensionality reduction methods have the advantage of finding manifold embedding owing to the highly nonlinear manifold of real-world data. However, they cannot be defined everywhere.
Deep machine learning for structural health monitoring on ship hulls using acoustic emission method
Published in Ships and Offshore Structures, 2021
Petros Karvelis, George Georgoulas, Vassilios Kappatos, Chrysostomos Stylios
An intuitive explanation of its effectiveness as well as the need for their complementary properties can be provided by projecting the multidimensional data (200 features) in a three-dimensional space. The ‘projection’ was created using a variation of the Stochastic Neighbour Embedding (SNE) method Hinton and Roweis (2002) called t-Distributed Stochastic Neighbour Embedding, t-SNE Maaten and Hinton (2008). Actually, t-SNE is a non linear dimensionality reduction technique which acts in a two-stage process: First, it starts by converting the high-dimensional Euclidean distances between data points into conditional probabilities that represent similarities in the high space. At the second stage, a probability distribution over the points in the low-dimensional map is defined and the method tries to minimise distance between the two distributions.