Trung Le

An optimal sphere and two large margins approach for novelty detection

We introduce a new model to deal with imbalanced data sets for novelty detection problems where the normal class of training data set can be majority or minority class. The key idea is to construct an optimal hypersphere such that the inside margin between the surface of this sphere and the normal data and the outside margin between that surface and the abnormal data are as large as possible. Depending on a specific real application of novelty detection, the two margins can be adjusted to achieve the best true positive and false positive rates. Experimental results on a number of data sets showed that the proposed model can provide better performance comparing with current models for novelty detection.

A theoretical framework for multi-sphere support vector data description

In support vector data description (SVDD) a spherically shaped boundary around a normal data set is used to separate this set from abnormal data. The volume of this data description is minimized to reduce the chance of accepting abnormal data. However the SVDD does not guarantee that the single spherically shaped boundary can best describe the normal data set if there are some distinctive data distributions in this set. A better description is the use of multiple spheres, however there is currently no investigation available. In this paper, we propose a theoretical framework to multi-sphere SVDD in which an optimisation problem and an iterative algorithm are proposed to determine model parameters for multi-sphere SVDD to provide a better data description to the normal data set. We prove that the classification error will be reduced after each iteration in this learning process. Experimental results on 28 well-known data sets show that the proposed multisphere SVDD provides lower classification error rate comparing with the standard single-sphere SVDD.

Fuzzy Multi-Sphere Support Vector Data Description

Current well-known data description methods such as Support Vector Data Description and Small Sphere Large Margin are conducted with assumption that data samples of a class in feature space are drawn from a single distribution. Based on this assumption, a single hypersphere is constructed to provide a good data description for the data. However, real-world data samples may be drawn from some distinctive distributions and hence it does not guarantee that a single hypersphere can offer the best data description. In this paper, we introduce a Fuzzy Multi-sphere Support Vector Data Description approach to address this issue. We propose to use a set of hyperspheres to provide a better data description for a given data set. Calculations for determining optimal hyperspheres and experimental results for applying this proposed approach to classification problems are presented.

Proximity multi-sphere support vector clustering

Support vector data description constructs an optimal hypersphere in feature space as a description of a data set. This hypersphere when mapped back to input space becomes a set of contours, and support vector clustering (SVC) employs these contours as cluster boundaries to detect clusters in the data set. However real-world data sets may have some distinctive distributions and hence a single hypersphere cannot be the best description. As a result, the set of contours in input space does not always detect all clusters in the data set. Another issue in SVC is that in some cases, it cannot preserve proximity notation which is crucial for cluster analysis, that is, two data points that are close to each other can be assigned to different clusters using cluster labelling method of SVC. To overcome these drawbacks, we propose Proximity Multi-sphere Support Vector Clustering which employs a set of hyperspheres to provide a better data description for data sets having distinctive distributions and a proximity graph to favour the proximity notation. Experimental results on different data sets are presented to evaluate the proposed clustering technique and compare it with SVC and other clustering techniques.

EEG-Based Person Verification Using Multi-Sphere SVDD and UBM

The use of brain-wave patterns extracted from electroencephalography (EEG) brain signals for person verification has been investigated recently. The challenge is that the EEG signals are noisy due to low conductivity of the human skull and the EEG data have unknown distribution. We propose a multi-sphere support vector data description (MSSVDD) method to reduce noise and to provide a mixture of hyperspheres that can describe the EEG data distribution. We also propose a MSSVDD universal background model (UBM) to model impostors in person verification. Experimental results show that our proposed methods achieved lower verification error rates than other verification methods.

Multiple distribution data description learning method for novelty detection

Current data description learning methods for novelty detection such as support vector data description and small sphere with large margin construct a spherically shaped boundary around a normal data set to separate this set from abnormal data. The volume of this sphere is minimized to reduce the chance of accepting abnormal data. However those learning methods do not guarantee that the single spherically shaped boundary can best describe the normal data set if there exist some distinctive data distributions in this set. We propose in this paper a new data description learning method that constructs a set of spherically shaped boundaries to provide a better data description to the normal data set. An optimisation problem is proposed and solving this problem results in an iterative learning algorithm to determine the set of spherically shaped boundaries. We prove that the classification error will be reduced after each iteration in our learning method. Experimental results on 23 well-known data sets show that the proposed method provides lower classification error rates.

Multiple distribution data description learning algorithm for novelty detection

Current data description learning methods for novelty detection such as support vector data description and small sphere with large margin construct a spherically shaped boundary around a normal data set to separate this set from abnormal data. The volume of this sphere is minimized to reduce the chance of accepting abnormal data. However those learning methods do not guarantee that the single spherically shaped boundary can best describe the normal data set if there exist some distinctive data distributions in this set. We propose in this paper a new data description learning method that constructs a set of spherically shaped boundaries to provide a better data description to the normal data set. An optimisation problem is proposed and solving this problem results in an iterative learning algorithm to determine the set of spherically shaped boundaries. We prove that the classification error will be reduced after each iteration in our learning method. Experimental results on 28 well-known data sets show that the proposed method provides lower classification error rates.

Fuzzy Multi-sphere Support Vector Data Description

Multi-sphere Support Vector Data Description (MS-SVDD) has been proposed in our previous work. MS-SVDD aims to build a set of spherically shaped boundaries that provide a better data description to the normal dataset and an iterative learning algorithm that determines the set of spherically shaped boundaries. MS-SVDD could improve classification rate for one-class classification problems comparing with SVDD. However MS-SVDD requires a small abnormal data set to build the spherically shaped boundaries for the normal data set. In this paper, we propose a new fuzzy MS-SVDD that can be used when only the normal data set is available. Experimental results on 14 well-known datasets and a comparison between fuzzy MS-SVDD and SVDD are also presented.

Multi-sphere support vector data description for brain-computer interface

Support vector data description (SVDD) has been widely used in pattern classification, however it does not provide high performance in brain-computer interface (BCI) classification problems since brain signals are noisy and chaotic. Brain data have distinct distributions and hence a hyper-sphere in SVDD could not well describe the data. We propose in this paper a multi-sphere approach to SVDD to have a better description for the brain data. We also propose a fuzzy clustering approach to optimize SVDD parameters. Experiments on the brain data set III for motor imagery problem in BCI Competition II were conducted to compare performance of SVDD and multi-sphere SVDD.

A new support vector machine method for medical image classification

One of the important problems in medical imaging is two-class classification, for example determination of benign from malignant cases in breast cancer treatment. In this paper we present a new support vector machine method for two-class medical image classification. The key idea of this method is to construct an optimal hypersphere such that both the interior margin between the surface of this sphere and the normal data, and the exterior margin between this surface and the abnormal data are as large as possible. The proposed method is easily implemented and can reduce both false positive and false negative error rates to obtain very good classification results. Experiments were performed on three medical image data sets to evaluate the proposed method.