Kernel density classification and boosting: an L2 analysis

Marzio, Marco Di; Taylor, Charles

doi:10.1007/s11222-005-6203-8

Cited by 24 publications

(10 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This conclusion is consistent with that found by Di Marzio and Taylor [5,6] , where boosting kernels gives higher order bias for both density estimation and classification. However, note that the current result uses L 2 Boosting for regression, rather than the Adaboost-like algorithms used in classification and density estimation.…”

Section: Boostnw Reduces the Bias Of The N-w Estimatorsupporting

confidence: 92%

“…In all three domains, methods exist which make use of a kernel function (kernel density estimation, kernel classifiers and kernel regression); these are often referred to as simply "nonparametric". Making use of these kernel methods, Di Marzio and Taylor [5][6][7] have indicated how boosting derives its success: namely, by reducing the bias of the estimators, with only moderate increases in variance. Using this result, one is able to use larger smoothing parameters and improve the overall quality of the final estimate.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Boosted Regression Estimates of Spatial Data: Pointwise Inference

Marzio¹,

Taylor²

2005

J. of Mathematics and Statistics

View full text Add to dashboard Cite

Abstract:In this study simple nonparametric techniques have been adopted to estimate the trend surface of the Swiss rainfall data. In particular we employed the Nadaraya-Watson smoother and in addition, an adapted-by boosting-version of it. Additionally, we have explored the use of the Nadaraya-Watson estimator for the construction of pointwise confidence intervals. Overall, boosting does seem to improve the estimate as much as previous examples and the results indicate that crossvalidation can be successfully used for parameter selection on real datasets. In addition, our estimators compare favorably with most of the techniques previously used on this dataset.

show abstract

Section: Boostnw Reduces the Bias Of The N-w Estimatorsupporting

confidence: 92%

Section: Introductionmentioning

confidence: 99%

Boosted Regression Estimates of Spatial Data: Pointwise Inference

Marzio¹,

Taylor²

2005

J. of Mathematics and Statistics

View full text Add to dashboard Cite

show abstract

“…The central idea of the KDE method is to apply kernel density estimation [8] on the BMD-list to estimate the probability density functions of the target class and the non-target classes, and then set the distance threshold δ so that at every point in the range of [0, δ] the probability density of belonging to the target class passes a probability threshold.…”

Section: Kde: Learning Distance Thresholds Using Kernel Density Estimmentioning

confidence: 99%

“…where K is a kernel function and h is a smoothing factor [8]. In this paper, we adopt the Gaussian kernel which has been popularly used.…”

Section: Kde: Learning Distance Thresholds Using Kernel Density Estimmentioning

confidence: 99%

Extracting Interpretable Features for Early Classification on Time Series

Xing¹,

Pei²,

Yu³

et al. 2011

Proceedings of the 2011 SIAM International Conference on Data Mining

143

145

View full text Add to dashboard Cite

Early classification on time series data has been found highly useful in a few important applications, such as medical and health informatics, industry production management, safety and security management. While some classifiers have been proposed to achieve good earliness in classification, the interpretability of early classification remains largely an open problem. Without interpretable features, application domain experts such as medical doctors may be reluctant to adopt early classification. In this paper, we tackle the problem of extracting interpretable features on time series for early classification. Specifically, we advocate local shapelets as features, which are segments of time series remaining in the same space of the input data and thus are highly interpretable. We extract local shapelets distinctly manifesting a target class locally and early so that they are effective for early classification. Our experimental results on seven benchmark real data sets clearly show that the local shapelets extracted by our methods are highly interpretable and can achieve effective early classification.

show abstract

“…Classification is performed with supervised non-parametric KDA [14,15]. The KDA is binary classifier, it simply give a `yes` or `no` answer to indicate whether a particular sample belong to a particular class or not.…”

Section: Classification Backgroundmentioning

confidence: 99%

An Analysis on Mammograms using KDA in Multi Transform Domain

Prathibha¹

2013

IJCA

View full text Add to dashboard Cite

Developing a robust Computer Aided Diagnosis (CADx) system for mammograms analysis has been the challenging task for years. The slight difference in X-ray attenuation between normal and abnormal glandular tissues makes the mammogram diagnosis complicated. The proposed CAD system discriminates the abnormal severity of mammograms into normal-benign (non-cancerous, non-spreadable), normalmalign (cancerous) ans benign-,malign, using wavelet features combined with spectral domain. Classification is performed on 206 different mammogram images from Mias database using Kernel Discriminant Analysis (KDA). KDA affords a non-parametric statistical approach with parzen window density estimation to estimate density function from a given sample dataset. The study reveals that the optimal smoothing parameters are increasing functions of the sample size of the complementary classes, features used to classify and value of the bandwidth.

show abstract

Kernel density classification and boosting: an L2 analysis

Cited by 24 publications

References 23 publications

Boosted Regression Estimates of Spatial Data: Pointwise Inference

Boosted Regression Estimates of Spatial Data: Pointwise Inference

Extracting Interpretable Features for Early Classification on Time Series

An Analysis on Mammograms using KDA in Multi Transform Domain

Contact Info

Product

Resources

About