A Note on the Generalization Performance of Kernel Classifiers with Margin

We present distribution independent bounds on the generalization misclassification performance of a family of kernel classifiers with margin. Support Vector Machine classifiers (SVM) stem out of this class of machines. The bounds are derived through computations of the $V_gamma$ dimension of a family of loss functions where the SVM one belongs to. Bounds that use functions of margin distributions (i.e. functions of the slack variables of SVM) are derived.

Kernel classifier construction using orthogonal forward selection and boosting with Fisher ratio class separability measure

A greedy technique is proposed to construct parsimonious kernel classifiers using the orthogonal forward selection method and boosting based on Fisher ratio for class separability measure. Unlike most kernel classification methods, which restrict kernel means to the training input data and use a fixed common variance for all the kernel terms, the proposed technique can tune both the mean vector and diagonal covariance matrix of individual kernel by incrementally maximizing Fisher ratio for class separability measure. An efficient weighted optimization method is developed based on boosting to append kernels one by one in an orthogonal forward selection procedure. Experimental results obtained using this construction technique demonstrate that it offers a viable alternative to the existing state-of-the-art kernel modeling methods for constructing sparse Gaussian radial basis function network classifiers. that generalize well.

Fast kernel classifier construction using orthogonal forward selection to minimise leave-one-out misclassification rate

We propose a simple yet computationally efficient construction algorithm for two-class kernel classifiers. In order to optimise classifier's generalisation capability, an orthogonal forward selection procedure is used to select kernels one by one by minimising the leave-one-out (LOO) misclassification rate directly. It is shown that the computation of the LOO misclassification rate is very efficient owing to orthogonalisation. Examples are used to demonstrate that the proposed algorithm is a viable alternative to construct sparse two-class kernel classifiers in terms of performance and computational efficiency.

Discriminative classifiers for speaker recognition

Speaker Recognition, Speaker Verification, Sparse Kernel Logistic Regression, Support Vector Machine

Domain Adaptation in remote sensing: increasing the portability of land-cover classifiers

Among the types of remote sensing acquisitions, optical images are certainly one of the most widely relied upon data sources for Earth observation. They provide detailed measurements of the electromagnetic radiation reflected or emitted by each pixel in the scene. Through a process termed supervised land-cover classification, this allows to automatically yet accurately distinguish objects at the surface of our planet. In this respect, when producing a land-cover map of the surveyed area, the availability of training examples representative of each thematic class is crucial for the success of the classification procedure. However, in real applications, due to several constraints on the sample collection process, labeled pixels are usually scarce. When analyzing an image for which those key samples are unavailable, a viable solution consists in resorting to the ground truth data of other previously acquired images. This option is attractive but several factors such as atmospheric, ground and acquisition conditions can cause radiometric differences between the images, hindering therefore the transfer of knowledge from one image to another. The goal of this Thesis is to supply remote sensing image analysts with suitable processing techniques to ensure a robust portability of the classification models across different images. The ultimate purpose is to map the land-cover classes over large spatial and temporal extents with minimal ground information. To overcome, or simply quantify, the observed shifts in the statistical distribution of the spectra of the materials, we study four approaches issued from the field of machine learning. First, we propose a strategy to intelligently sample the image of interest to collect the labels only in correspondence of the most useful pixels. This iterative routine is based on a constant evaluation of the pertinence to the new image of the initial training data actually belonging to a different image. Second, an approach to reduce the radiometric differences among the images by projecting the respective pixels in a common new data space is presented. We analyze a kernel-based feature extraction framework suited for such problems, showing that, after this relative normalization, the cross-image generalization abilities of a classifier are highly increased. Third, we test a new data-driven measure of distance between probability distributions to assess the distortions caused by differences in the acquisition geometry affecting series of multi-angle images. Also, we gauge the portability of classification models through the sequences. In both exercises, the efficacy of classic physically- and statistically-based normalization methods is discussed. Finally, we explore a new family of approaches based on sparse representations of the samples to reciprocally convert the data space of two images. The projection function bridging the images allows a synthesis of new pixels with more similar characteristics ultimately facilitating the land-cover mapping across images.

Kernel-based mapping of orographic rainfall enhancement in the Swiss Alps as detected by weatherradar

In this paper, we develop a data-driven methodology to characterize the likelihood of orographic precipitation enhancement using sequences of weather radar images and a digital elevation model (DEM). Geographical locations with topographic characteristics favorable to enforce repeatable and persistent orographic precipitation such as stationary cells, upslope rainfall enhancement, and repeated convective initiation are detected by analyzing the spatial distribution of a set of precipitation cells extracted from radar imagery. Topographic features such as terrain convexity and gradients computed from the DEM at multiple spatial scales as well as velocity fields estimated from sequences of weather radar images are used as explanatory factors to describe the occurrence of localized precipitation enhancement. The latter is represented as a binary process by defining a threshold on the number of cell occurrences at particular locations. Both two-class and one-class support vector machine classifiers are tested to separate the presumed orographic cells from the nonorographic ones in the space of contributing topographic and flow features. Site-based validation is carried out to estimate realistic generalization skills of the obtained spatial prediction models. Due to the high class separability, the decision function of the classifiers can be interpreted as a likelihood or susceptibility of orographic precipitation enhancement. The developed approach can serve as a basis for refining radar-based quantitative precipitation estimates and short-term forecasts or for generating stochastic precipitation ensembles conditioned on the local topography.

Comparing Support Vector Machines with Gaussian Kernels to Radial Basis Function Classifiers

The Support Vector (SV) machine is a novel type of learning machine, based on statistical learning theory, which contains polynomial classifiers, neural networks, and radial basis function (RBF) networks as special cases. In the RBF case, the SV algorithm automatically determines centers, weights and threshold such as to minimize an upper bound on the expected test error. The present study is devoted to an experimental comparison of these machines with a classical approach, where the centers are determined by $k$--means clustering and the weights are found using error backpropagation. We consider three machines, namely a classical RBF machine, an SV machine with Gaussian kernel, and a hybrid system with the centers determined by the SV method and the weights trained by error backpropagation. Our results show that on the US postal service database of handwritten digits, the SV machine achieves the highest test accuracy, followed by the hybrid approach. The SV approach is thus not only theoretically well--founded, but also superior in a practical application.

On the annealed VC entropy for margin classifiers: A statistical mechanics study

Using techniques from Statistical Physics, the annealed VC entropy for hyperplanes in high dimensional spaces is calculated as a function of the margin for a spherical Gaussian distribution of inputs.

Dynamic classifiers for neonatal brain monitoring

Brain injury due to lack of oxygen or impaired blood flow around the time of birth, may cause long term neurological dysfunction or death in severe cases. The treatments need to be initiated as soon as possible and tailored according to the nature of the injury to achieve best outcomes. The Electroencephalogram (EEG) currently provides the best insight into neurological activities. However, its interpretation presents formidable challenge for the neurophsiologists. Moreover, such expertise is not widely available particularly around the clock in a typical busy Neonatal Intensive Care Unit (NICU). Therefore, an automated computerized system for detecting and grading the severity of brain injuries could be of great help for medical staff to diagnose and then initiate on-time treatments. In this study, automated systems for detection of neonatal seizures and grading the severity of Hypoxic-Ischemic Encephalopathy (HIE) using EEG and Heart Rate (HR) signals are presented. It is well known that there is a lot of contextual and temporal information present in the EEG and HR signals if examined at longer time scale. The systems developed in the past, exploited this information either at very early stage of the system without any intelligent block or at very later stage where presence of such information is much reduced. This work has particularly focused on the development of a system that can incorporate the contextual information at the middle (classifier) level. This is achieved by using dynamic classifiers that are able to process the sequences of feature vectors rather than only one feature vector at a time.

Sensitivity and specificity of machine learning classifiers for glaucoma diagnosis using Spectral Domain OCT and standard automated perimetry

PURPOSE: To evaluate the sensitivity and specificity of machine learning classifiers (MLCs) for glaucoma diagnosis using Spectral Domain OCT (SD-OCT) and standard automated perimetry (SAP). METHODS: Observational cross-sectional study. Sixty two glaucoma patients and 48 healthy individuals were included. All patients underwent a complete ophthalmologic examination, achromatic standard automated perimetry (SAP) and retinal nerve fiber layer (RNFL) imaging with SD-OCT (Cirrus HD-OCT; Carl Zeiss Meditec Inc., Dublin, California). Receiver operating characteristic (ROC) curves were obtained for all SD-OCT parameters and global indices of SAP. Subsequently, the following MLCs were tested using parameters from the SD-OCT and SAP: Bagging (BAG), Naive-Bayes (NB), Multilayer Perceptron (MLP), Radial Basis Function (RBF), Random Forest (RAN), Ensemble Selection (ENS), Classification Tree (CTREE), Ada Boost M1(ADA),Support Vector Machine Linear (SVML) and Support Vector Machine Gaussian (SVMG). Areas under the receiver operating characteristic curves (aROC) obtained for isolated SAP and OCT parameters were compared with MLCs using OCT+SAP data. RESULTS: Combining OCT and SAP data, MLCs' aROCs varied from 0.777(CTREE) to 0.946 (RAN).The best OCT+SAP aROC obtained with RAN (0.946) was significantly larger the best single OCT parameter (p<0.05), but was not significantly different from the aROC obtained with the best single SAP parameter (p=0.19). CONCLUSION: Machine learning classifiers trained on OCT and SAP data can successfully discriminate between healthy and glaucomatous eyes. The combination of OCT and SAP measurements improved the diagnostic accuracy compared with OCT data alone.

DECISION TREE CLASSIFIERS FOR STAR/GALAXY SEPARATION

We study the star/galaxy classification efficiency of 13 different decision tree algorithms applied to photometric objects in the Sloan Digital Sky Survey Data Release Seven (SDSS-DR7). Each algorithm is defined by a set of parameters which, when varied, produce different final classification trees. We extensively explore the parameter space of each algorithm, using the set of 884,126 SDSS objects with spectroscopic data as the training set. The efficiency of star-galaxy separation is measured using the completeness function. We find that the Functional Tree algorithm (FT) yields the best results as measured by the mean completeness in two magnitude intervals: 14 <= r <= 21 (85.2%) and r >= 19 (82.1%). We compare the performance of the tree generated with the optimal FT configuration to the classifications provided by the SDSS parametric classifier, 2DPHOT, and Ball et al. We find that our FT classifier is comparable to or better in completeness over the full magnitude range 15 <= r <= 21, with much lower contamination than all but the Ball et al. classifier. At the faintest magnitudes (r > 19), our classifier is the only one that maintains high completeness (> 80%) while simultaneously achieving low contamination (similar to 2.5%). We also examine the SDSS parametric classifier (psfMag - modelMag) to see if the dividing line between stars and galaxies can be adjusted to improve the classifier. We find that currently stars in close pairs are often misclassified as galaxies, and suggest a new cut to improve the classifier. Finally, we apply our FT classifier to separate stars from galaxies in the full set of 69,545,326 SDSS photometric objects in the magnitude range 14 <= r <= 21.

Kernel machines for epilepsy diagnosis via EEG signal classification: A comparative study

Objective: We carry out a systematic assessment on a suite of kernel-based learning machines while coping with the task of epilepsy diagnosis through automatic electroencephalogram (EEG) signal classification. Methods and materials: The kernel machines investigated include the standard support vector machine (SVM), the least squares SVM, the Lagrangian SVM, the smooth SVM, the proximal SVM, and the relevance vector machine. An extensive series of experiments was conducted on publicly available data, whose clinical EEG recordings were obtained from five normal subjects and five epileptic patients. The performance levels delivered by the different kernel machines are contrasted in terms of the criteria of predictive accuracy, sensitivity to the kernel function/parameter value, and sensitivity to the type of features extracted from the signal. For this purpose, 26 values for the kernel parameter (radius) of two well-known kernel functions (namely. Gaussian and exponential radial basis functions) were considered as well as 21 types of features extracted from the EEG signal, including statistical values derived from the discrete wavelet transform, Lyapunov exponents, and combinations thereof. Results: We first quantitatively assess the impact of the choice of the wavelet basis on the quality of the features extracted. Four wavelet basis functions were considered in this study. Then, we provide the average accuracy (i.e., cross-validation error) values delivered by 252 kernel machine configurations; in particular, 40%/35% of the best-calibrated models of the standard and least squares SVMs reached 100% accuracy rate for the two kernel functions considered. Moreover, we show the sensitivity profiles exhibited by a large sample of the configurations whereby one can visually inspect their levels of sensitiveness to the type of feature and to the kernel function/parameter value. Conclusions: Overall, the results evidence that all kernel machines are competitive in terms of accuracy, with the standard and least squares SVMs prevailing more consistently. Moreover, the choice of the kernel function and parameter value as well as the choice of the feature extractor are critical decisions to be taken, albeit the choice of the wavelet family seems not to be so relevant. Also, the statistical values calculated over the Lyapunov exponents were good sources of signal representation, but not as informative as their wavelet counterparts. Finally, a typical sensitivity profile has emerged among all types of machines, involving some regions of stability separated by zones of sharp variation, with some kernel parameter values clearly associated with better accuracy rates (zones of optimality). (C) 2011 Elsevier B.V. All rights reserved.

Extensions of discrete triangular distributions and boundary bias in kernel estimation for discrete functions

Asymmetric discrete triangular distributions are introduced in order to extend the symmetric ones serving for discrete associated kernels in the nonparametric estimation for discrete functions. The extension from one to two orders around the mode provides a large family of discrete distributions having a finite support. Establishing a bridge between Dirac and discrete uniform distributions, some different shapes are also obtained and their properties are investigated. In particular, the mean and variance are pointed out. Applications to discrete kernel estimators are given with a solution to a boundary bias problem. (C) 2010 Elsevier B.V. All rights reserved.

Meat Quality of Lambs Fed on Palm Kernel Meal, a By-product of Biodiesel Production

This study aimed to establish the optimum level of palm kernel meal in the diet of Santa Ines lambs based on the sensorial characteristics and fatty acid profile of the meat. We used 32 lambs with a starting age of 4 to 6 months and mean weight of 22 2.75 kg, kept in individual stalls. The animals were fed with Tifton-85 hay and a concentrate mixed with 0.0, 6.5, 13.0 or 19.5% of palm kernel meal based on the dry mass of the complete diet. These levels formed the treatments. Confinement lasted 80 days and on the last day the animals were fasted and slaughtered. After slaughter, carcasses were weighed and sectioned longitudinally, along the median line, into two antimeres. Half-carcasses were then sliced between the 12th and 13th ribs to collect the loin (longissimus dorsi), which was used to determine the sensorial characteristics and fatty acid profile of the meat. For sensorial evaluation, samples of meat were given to 54 judges who evaluated the tenderness, juiciness, appearance, aroma and flavor of the meat using a hedonic scale. Fatty acids were determined by gas chromatography. The addition of palm kernel meal to the diet had no effect on the sensorial characteristics of meat juiciness, appearance, aroma or flavor. However, tenderness showed a quadratic relationship with the addition of the meal to the diet. The concentration of fatty acids C12:0, C14:0 and C16:0 increased with the addition of palm kernel meal, as did the sum of medium-chain fatty acids and the atherogenicity index. Up to of 19.5% of the diet of Santa Ines lambs can be made up of palm kernel meal without causing significant changes in sensorial characteristics. However, the fatty acid profile of the meat was altered.