Experimentally optimal ν in support vector regression for different noise models and parameter settings

Chalimourda, Athanassia; Schölkopf, Bernhard; Smola, Alex

doi:10.1016/s0893-6080(03)00209-0

Cited by 124 publications

(50 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We used a v-SVR (Scholkopf et al, 1999) with v set to 1, since a linear kernel was used (see below). The other parameter of the SVR model C was set to the maximum t2 VSWM times the number of subjects, according to what is recommended in the literature (Chalimourda et al, 2004). Both of the parameters of the SVR model were therefore not fitted to the dataset but determined according to the usual criteria.…”

Section: Support Vector Regression Model and Analysismentioning

confidence: 99%

See 1 more Smart Citation

Structural Maturation and Brain Activity Predict Future Working Memory Capacity during Childhood Development

2014

View full text Add to dashboard Cite

Human working memory capacity develops during childhood and is a strong predictor of future academic performance, in particular, achievements in mathematics and reading. Predicting working memory development is important for the early identification of children at risk for poor cognitive and academic development. Here we show that structural and functional magnetic resonance imaging data explain variance in children's working memory capacity 2 years later, which was unique variance in addition to that predicted using cognitive tests. While current working memory capacity correlated with frontoparietal cortical activity, the future capacity could be inferred from structure and activity in basal ganglia and thalamus. This gives a novel insight into the neural mechanisms of childhood development and supports the idea that neuroimaging can have a unique role in predicting children's cognitive development.

show abstract

Section: Support Vector Regression Model and Analysismentioning

confidence: 99%

“…Several studies have used MRI to show developmental changes in working memory (WM)-related brain activation Kwon et al, 2002;Crone et al, 2006;Scherf et al, 2006) but not to predict future WM capacity. Machine learning algorithms, based on multivariate data, are promising for this goal (Haxby et al, 2001).…”

Section: Introductionmentioning

confidence: 99%

Structural Maturation and Brain Activity Predict Future Working Memory Capacity during Childhood Development

2014

View full text Add to dashboard Cite

show abstract

“…Currently, SVM algorithm has been widely applied in pattern recognition, regression estimation, probability density estimation, dimensionality reduction, etc. [16][17][18]. The evaluation of circular economy for iron and steel enterprise is a typical non-linear, high-dimensional pattern classification problem, as well as a multi-link, multi-level system.…”

Section: Support Vector Machinementioning

confidence: 99%

On Evaluation Model of Circular Economy for Iron and Steel Enterprise Based on Support Vector Machines with Heuristic Algorithm for Tuning Hyper-parameters

Zhou¹,

Chen²,

Xiao³

2013

Appl. Math. Inf. Sci.

View full text Add to dashboard Cite

With more severe resource scarcity and environmental problems, the evaluation of circular economy in microcosmic level has become the focus of the academic world. Based on the concept of circular economy, this paper not only structures the evaluation index system of circular economy for iron and steel enterprises, builds the evaluation model of circular economy for iron and steel enterprises based on Support Vector Machine (SVM) with Radial Basis Function (RBF) kernel, but achieves the optimization of kernel function parameters, penalty factors and insensitive parameters based on a heuristic algorithm for tuning hyper-parameters. Furthermore, the evaluation model is tested for circular economy evaluation in the major iron and steel enterprises in China. The research demonstrates that the evaluation results of heuristic algorithm with SVM are more accurate, and this model is more suitable for iron and steel enterprises to evaluate circular economy, compared with the evaluation method of BP neural network.

show abstract

“…The signal to error (S/E) ratio was computed in decibels as (37) in the training set. Means and standard deviations of S/E were averaged over 100 realizations.…”

Section: B Training and Validation Signalsmentioning

confidence: 99%

“…However, several criteria are available in the SVM literature to tune the free parameters when little or no knowledge about the problem is available [33]- [37]. Note that, in general, and can be optimized by using the same methodology as in [34], but such an analysis is beyond the scope of the present paper.…”

Section: Tuning the Free Parametersmentioning

confidence: 99%

Nonuniform Interpolation of Noisy Signals Using Support Vector Machines

Rojo-Álvarez

Figuera

Martinez-Cruz

et al. 2007

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

Experimentally optimal ν in support vector regression for different noise models and parameter settings

Cited by 124 publications

References 3 publications

Structural Maturation and Brain Activity Predict Future Working Memory Capacity during Childhood Development

Structural Maturation and Brain Activity Predict Future Working Memory Capacity during Childhood Development

On Evaluation Model of Circular Economy for Iron and Steel Enterprise Based on Support Vector Machines with Heuristic Algorithm for Tuning Hyper-parameters

Nonuniform Interpolation of Noisy Signals Using Support Vector Machines

Contact Info

Product

Resources

About