D. M. Tax and R. P. Duin, "Support Vector Data Description," Machine Learning 55, pp. 45--66, 2004. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Probability Density Estimation from Optimally Condensed Data.. - Girolami, He (2003) (1 citation) (Correct)
.... exactly as d K h (x, x i )K h (x, x j )dx Denoting h (x, x i )K h (x, x j )dx by i , x j ) then the quadratic left hand term can be written as i , x j ) A similar constrained quadratic form has been utilised previously to obtain a minimum volume description of a data sample [32] or to obtain a sample estimate of the distribution support [30] where it has been observed empirically that the extremal points in the sample are given a finite weighting coefficient. This can be viewed as placing finite weight to points in regions of low density which is in contrast to the ....
....shown. points lie in the centre of the distribution and the shape they form is somewhat reminiscent of that obtained by Principal Curves [8] This similarity may form an interesting area of future investigation. This observation is in contrast to the support vector data description methods [30] [32] where the boundary points of the sample tend to be selected. C. Comparative Experiments The first experiment in this section compares the RSDE with the SVM approach to density estimation [19] The 1 D density function employed in [19] is used in this experiment i.e. p(x) 2 # 2# exp( 0.5 x ....
Tax, D.M.J. and Duin, R.P.W. (1999) Support Vector Data Description, Pattern Recognition Letters, 20:(1113) , pp 1191-1199.
One-Class LP Classifier for Dissimilarity - Representations Elzbieta Pekalska Self-citation (Tax Duin) (Correct)
No context found.
D.M.J. Tax and R.P.W. Duin. Support vector data description. Machine Learning, 2002. accepted.
One-class LP classifier for dissimilarity - Representations Elzbieta Pekalska Self-citation (Tax Duin) (Correct)
No context found.
D.M.J. Tax and R.P.W. Duin. Support vector data description. Machine Learning, 2002. accepted.
Feature Scaling in Support Vector Data Description - Duin Self-citation (Tax Duin) (Correct)
....The efficiency of appropriate rescaling the feature space is shown on a artificial datasets and a real world handwritten digits dataset. The results are presented in the section Experiments. 2 Theory Now we would like to give a short description of the SVDD. For more information we refer to [2, 1]. In the SVDD the data is enclosed by a hypersphere with minimum volume. By minimizing the volume of the feature space, or equivalently minimizing the radius R we hope to minimize the chance of accepting outlier objects. So in analogy to the support vector classifier [4] we can define the ....
....to evaluate how well the rescaling procedures work on an artificial and a realworld data. We start with describing artificial data. Higleyman (a normally distributed two dimension data with different covariance matrices Highleyman classes are defined by N( 1 1] 1 0; 0 0. 25] for class A and N([2 0], 0.01 0; 0 4] for class B) difficult (a 0.51 (0.06) 0.38 (0.06) difficult 2D 0.80 (0.21) 0.18 (0.07) difficult 5D 0.93 (0.12) 0.30 (0.07) difficult 10D 0.93 (0.11) 0.42 (0.07) banana 0.38 (0.05) 0.04 (0.04) for SVDD with the error of the first kind column and 0.1 in the second. ....
D.M.J. Tax, R.P.W. Duin, 'Support Vector Data Description' , Pattern Recognition Letters, December 1999, vol. 20(11-13), pg. 1191-1199
Identification of Image Variations based on - Equivalence Classes Maret (Correct)
No context found.
D. M. Tax and R. P. Duin, "Support Vector Data Description," Machine Learning 55, pp. 45--66, 2004.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC