Locally linear embedding based on correntropy measure for visualization and classification

Daza-Santacoloma, Genaro; Castellanos-Domínguez, G.; Prı́ncipe, José C.

doi:10.1016/j.neucom.2011.09.015

Cited by 19 publications

(8 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [7] an extension of the LLE method is presented, which we called Class Label Locally Linear Embedding -CLLE. This technique employs class labels as extra information to guide the dimensionality reduction procedure, to preserve the local geometry of the data while providing a discriminative strategy during the mapping.…”

Section: Class Label Locally Linear Embedding -Cllementioning

confidence: 99%

See 1 more Smart Citation

Human Activity Recognition by Class Label LLE

Valencia-Aguirre

Álvarez-Meza

Daza-Santacoloma

et al. 2012

Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications

Self Cite

View full text Add to dashboard Cite

Human motion analysis has emerged as an important area of research for different fields and applications. However, analyzing image and video sequences to perform tasks such as action recognition, becomes a challenge due to the high dimensionality of this type of data, not mentioning the restrictions in the recording conditions (lighting, angle, distances, etc). In that sense, we propose a framework for human action recognition, which involves a preprocessing stage that decreases the influence of the record conditions in the analysis. Further, our proposal is based on a new supervised feature extraction technique that includes class label information in the mapping process, to enhance both the underlying data structure unfolding and the margin of separability among classes. Proposed methodology is tested on a benchmark dataset. Attained results show how our approach obtains a suitable performance using straightforward classifiers.

show abstract

Section: Class Label Locally Linear Embedding -Cllementioning

confidence: 99%

“…Next, the reconstruction error e R and the margin μ can be computed as e R (β) = tr Y β MY β , and μ (β) = tr Y βM Y β . Looking forward the minimization of e R (β) and the maximization of μ (β), the parametric plot e R (β) vs μ (β) can be used to study the behavior of these quantities [7].…”

Section: Class Label Locally Linear Embedding -Cllementioning

confidence: 99%

Human Activity Recognition by Class Label LLE

Valencia-Aguirre

Álvarez-Meza

Daza-Santacoloma

et al. 2012

Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications

Self Cite

View full text Add to dashboard Cite

show abstract

“…Correntropy is a method to solve the problem of overall worsening of estimation results because of wrong values that are usually contained in least squares. The concept of correntropy is used in image analysis, 22 acoustic processing of voice, 23 and so on; in the field of power systems, there are studies on state estimation of transmission systems 24 and output forecasting of wind power generation. 25 Considering massive introduction of photovoltaic generation and other renewable energies, one has to take into account possibility of bad data that are cannot be classified as outliers.…”

Section: Introductionmentioning

confidence: 99%

Differential Evolutionary Particle Swarm Optimization for Load Adjustment Distribution State Estimation using Correntropy

Iwata

Fukuyama

2018

IEEJ Trans. PE

View full text Add to dashboard Cite

This paper proposes differential evolutionary particle swarm optimization (DEEPSO) for load adjustment distribution state estimation (DSE) using correntropy. Practical equipment in distribution systems causes nonlinear characteristics in an objective function and evolutionary computation methods have been applied to DSE so far. This paper applies DEEPSO in order to improve estimation equality. Minimization of sum of square errors by the least squares method has a problem when outliers exit in measured values. Quality of estimated results is largely affected by the outliers using the least squares method, while correntropy has a possibility not to be affected by the outliers. The proposed method is applied to a typical distribution system. The results indicate that the proposed DEEPSO-based method can improve estimation results compared with conventional particle swarm optimization (PSO) and hybrid PSO-based method, and the correntropy-based proposed method can estimate distribution system conditions more accurately than the conventional least squares method. K E Y W O R D Scorrentropy, differential evolutionary particle swarm optimization, distribution system, load adjustment distribution state estimation Electr Eng Jpn. 2018;205:11-21.

show abstract

“…The linear methods, such as Principal Component Analysis (PCA) [13] and Multidimensional Scaling (MDS) [14], are used when data sets have some linear subspace, and they use Euclidean distance to compute the distance among points [15]. Nonlinear methods dealing with nonlinear structures includes, Isomap [15], Curvilinear Distance Analysis (CDA) [16], Locally Linear Embedding (LLE) [17], Stochastic Proximity Embedding (SPE) [18] and Stochastic Neighbor Embedding (SNE) [19]. These methods attempt to inherit the relations among global neighborhood points to local corresponding points in projected space.…”

Section: Introductionmentioning

confidence: 99%