Clustering Based One-Class Classification for Compliance Verification of the Comprehensive Nuclear-Test-Ban Treaty

Abstract

Monitoring the levels of radioxenon isotopes in the atmosphere has been proposed as a means of verifying the Comprehensive Nuclear-Test-Ban Treaty (CTBT). This translates into a classification problem, whereby the measured concentrations either belong to an explosion class or a background class. Instances drawn from the explosions class are extremely rare, if not non-existent. Therefore, the resulting dataset is extremely imbalanced, and inherently suited for one-class classification. Further exacerbating the problem is the fact that the background distribution can be extremely complex, and thus, modelling it using one-class learning is difficult. In order to improve upon the previous classification results, we investigate the augmentation of one-class learning methods with clustering. The purpose of clustering is to convert a complex distribution into simpler distributions, the clusters, over which more effective models can be built. The resulting model, built from oneclass learners trained over the clusters, performs more effectively than a model that is built over the original distribution. This thesis is empirically tested on three different data domains; in particular, a number of artificial datasets, datasets from the UCI repository, and data modelled after the extremely challenging CTBT. The results offer credence to the fact that there is an improvement in performance when clustering is used with one-class classification on complex distributions.

Information-assisted volume rendering and visual evaluation through machine intelligence

“…For these reasons, one-class classification has found its application in several fields such as engine fault detection [104], medical diagnosis [105], nuclear testing [106], web page classification [107] and network intrusion detection [108].…”

Cluster validation in clustering‐based one‐class classification

“…Ensembles increase the classification performance because they can reduce the variance (Breiman, ) and the effect of individual classifiers' bias (Rokach, ). Different classifiers in the ensemble are usually created by using a different selection of objects, as is the case of Valentini and Dietterich's multiple ocSVM classifiers (Valentini & Dietterich, ), structured one‐class classification (Wang et al, ; Sharma et al, ), Bagging‐TPMiner (Medina‐Pérez et al, ), and Giacinto et al's modular ensemble (Giacinto et al, , ). Another popular alternative is to create the classifiers in the ensemble by using a different selection of attributes, for example, one attribute per classifier (Juszczak & Duin, ) or subsets of original or derived attributes (Nanni, ; Biggio et al, ; Cheplygina & Tax, ; Krawczyk, ; Rodríguez et al, ; Tax & Duin, ).…”

Optimal construction of one-against-one classifier based on meta-learning

“…The third strategy is to decompose the original problem into several one-class subproblems [22][23][24][25][26][27]. This strategy utilizes oneclass classification algorithms, such as parzen window, support vector data description, kernel principal component analysis, for multi-class classification.…”