Pattern Recognition with Optimal Margin Classifiers

Boser, Bernhard E.

doi:10.1007/978-3-642-78646-4_7

Cited by 3 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Boser et al, ( 1992) suggested to map input data into a high dimensional feature space through some nonlinear mapping. The transfonnation to a higher dimensional space spreads the data out in a way that facilitates the finding oflinear hyperplane.…”

Section: Support Vector Machinesmentioning

confidence: 99%

Kernel Methods in Remote Sensing: A Review

Pal¹

2009

ISH Journal of Hydraulic Engineering

View full text Add to dashboard Cite

Kernel-based machine learning algorithms are based on mapping data from the original input feature space to a kernel feature space of higher dimensionality to solve a linear problem in that space. Over the last decade, kernel based classification and regression approaches such as support vector machines have widely been used in remote sensing as well as in various civil engineering applications. In spite of their better performance with different datasets, support vector machines still suffer from shortcomings such as visualization/interpretation of model, choice of kernel and kernel specific parameter as well as the regularization parameter. Relevance vector machines are another kernel based approach being explored for classification and regression with in last few years. The advantages of the relevance vector machines over the support vector machines is the availability of probabilistic predictions, using arbitrary kernel functions and not requiring setting of the regularization parameter. This paper presents a state-of-the-art review of SVM and RVM in remote sensing and provides some details of their use in other civil engineering application also.

show abstract

Section: Support Vector Machinesmentioning

confidence: 99%

Kernel Methods in Remote Sensing: A Review

Pal¹

2009

ISH Journal of Hydraulic Engineering

View full text Add to dashboard Cite

show abstract

“…Support vector machines have been extended to classifying a feature vector among more than two diagnostic classes and for computing nonlinear separating planes by first using nonlinear mapping to transfer the features to a higher dimensional space and then estimating an optimal hyperplane in the higher dimensional space. 45,46 The accuracy of an SVM depends upon the optimal mapping to the higher dimensional space, with the consequence that the computational complexity of defining a classifier is high. 47 Furthermore, because the decision boundary learned by an SVM is defined by the support vectors only, the classification boundary is highly sensitive to noise in these vectors.…”

Section: Methodsmentioning

confidence: 99%

Using Copula distributions to support more accurate imaging-based diagnostic classifiers for neuropsychiatric disorders

Bansal

Hao

Liu

et al. 2014

Magnetic Resonance Imaging

View full text Add to dashboard Cite

Many investigators have tried to apply machine learning techniques to magnetic resonance images (MRIs) of the brain in order to diagnose neuropsychiatric disorders. Usually the number of brain imaging measures (such as measures of cortical thickness and measures of local surface morphology) derived from the MRIs (i.e., their dimensionality) has been large (e.g. >10) relative to the number of participants who provide the MRI data (<100). Sparse data in a high dimensional space increases the variability of the classification rules that machine learning algorithms generate, thereby limiting the validity, reproducibility, and generalizability of those classifiers. The accuracy and stability of the classifiers can improve significantly if the multivariate distributions of the imaging measures can be estimated accurately. To accurately estimate the multivariate distributions using sparse data, we propose to estimate first the univariate distributions of imaging data and then combine them using a Copula to generate more accurate estimates of their multivariate distributions. We then sample the estimated Copula distributions to generate dense sets of imaging measures and use those measures to train classifiers. We hypothesize that the dense sets of brain imaging measures will generate classifiers that are stable to variations in brain imaging measures, thereby improving the reproducibility, validity, and generalizability of diagnostic classification algorithms in imaging datasets from clinical populations. In our experiments, we used both computer-generated and real-world brain imaging datasets to assess the accuracy of multivariate Copula distributions in estimating the corresponding multivariate distributions of real-world imaging data. Our experiments showed that diagnostic classifiers generated using imaging measures sampled from the Copula were significantly more accurate and more reproducible than were the classifiers generated using either the real-world imaging measures or their multivariate Gaussian distributions. Thus, our findings demonstrate that estimated multivariate Copula distributions can generate dense sets of brain imaging measures that can in turn be used to train classifiers, and those classifiers are significantly more accurate and more reproducible than are those generated using real-world imaging measures alone.

show abstract

“…For classification purposes, a Support Vector Machine (SVM) is used [3]. In order to train and test an SVM model, it is first trained with a training set for which features and target labels are known.…”

Section: Classificationmentioning

confidence: 99%

Computer aided diagnosis of Alzheimers disease by automatically obtaining the best coronal slices for multi classification recognition

Carrillo¹,

Castillo²,

Rojas³

et al. 2018

Sixth International Conference on Advances in Bio-Informatics, Bio-Technology and Environmental Engineering - ABBE 2018

View full text Add to dashboard Cite

The goal of this contribution is to find out a set of Y slices (coronal slices) from MRIs of patients with Alzheimer's Disease (AD), Mild Cognitive Impairment (MCI), and Normal images, that provides the maximum accuracy in a multiclass classification system. Images are preprocessed and 2D wavelet coefficients are extracted to form a feature matrix. Using a feature selection algorithm called mRMR, the best features from the matrix are extracted; then, the dimension of the feature vectors is reduced using PCA and finally, it is used to train an SVM to perform multi-class classification. In order to find the best combinations of coronal slices, a multi-objective genetic optimization methodology based on NSGA-II is used and a set of different solutions are extracted from the Pareto front. More relevant solutions are selected using more flexible criteria than that of the Pareto front, and examine what slices and accuracies are achieved. The multi-classification accuracies obtained by the proposed method are 94.2% (the individual of the Pareto front with the highest accuracy) or 91.9% (using the best 13 slices according to their frequency of presence in the Pareto front). It is important to note that not only a good accuracy is obtained in the classification, but also new knowledge about the most relevant coronal slices to distinguish the four MRI classes (AD, Normal, MCI, LMCI).

show abstract

Pattern Recognition with Optimal Margin Classifiers

Cited by 3 publications

References 12 publications

Kernel Methods in Remote Sensing: A Review

Kernel Methods in Remote Sensing: A Review

Using Copula distributions to support more accurate imaging-based diagnostic classifiers for neuropsychiatric disorders

Computer aided diagnosis of Alzheimers disease by automatically obtaining the best coronal slices for multi classification recognition

Contact Info

Product

Resources

About