Restricted default theories: Expressive power and outlier detection tasks

The development of effective knowledge discovery techniques has become a very active research area in recent years due to the important impact it has had in several relevant application domains. One interesting task therein is that of singling out anomalous individuals from a given population, for example, to detect rare events in time-series analysis settings, or to identify objects whose behavior is deviant w.r.t. a codified standard set of rules. Such exceptional individuals are usually referred to as outliers in the literature. In this article, the concept of outlier is formally stated in the context of knowledge-based systems, by generalizing that originally proposed in Angiulli et al. [2003] in the context of default theories. The chosen formal framework here is that of logic programming, wherein potential applications of techniques for outlier detection are thoroughly discussed. The proposed formalization is a novel one and helps to shed light on the nature of outliers occurring in logic bases. Also the exploitation of minimality criteria in outlier detection is illustrated. The computational complexity of outlier detection problems arising in this novel setting is also thoroughly investigated and accounted for in the paper. Finally, rewriting algorithms are proposed that transform any outlier detection problem into an equivalent inference problem under stable model semantics, thereby making outlier computation effective and realizable on top of any stable model solver.

Developing species distribution models (SDMs) requires high‐quality species occurrence records. These records, stemming from various sources with different sampling procedures, are often archived in open‐access databases, making automated data quality checks inevitable. Temporal, geographic, and taxonomic quality checks are usually conducted in SDM workflows, but checking for records distant in environmental space, i.e. outliers, is often ignored. Here, we present ‘specleanr', an R package that contains 20 outlier detection methods (ODMs) that can be ensembled to identify potential outliers in environmental predictors. These methods are categorized into 1) species‐specific ecological range, 2) univariate, and 3) multivariate ODMs. All potential outliers flagged by the different methods are pooled to identify absolute outliers (records appearing in multiple methods). The local regression (LOESS) method is then used to automatically set a threshold that optimally identifies the absolute outliers. Additionally, clustering records into poor, fair, moderate, very strong, and perfect outliers, as well as non‐outliers, is possible based on each record's likelihood as a potential outlier, which allows expert assessment. We demonstrated the approach to 15 fish species from the Danube River Basin, including native, alien, threatened, and common species. We fitted SDMs using bioclimatic and hydromorphological parameters. We compared the model area under the curve (AUC) before and after outlier removal using three scenarios: 1) the LOESS method, 2) removing very strong outliers, and 3) removing perfect outliers. The results showed a significant improvement in the model AUC, with generally small to moderate effect sizes after outlier removal. ‘specleanr' is generalizable across taxonomic groups, data types, ecological realms, and geographic regions. Beyond SDMs, it can also be broadly used in general data analysis where outlier detection is essential. We provide detailed vignettes to support package use. ‘specleanr' offers a user‐friendly and reproducible approach for handling outliers in biogeographical modeling and general data analysis workflows.

Reiter's default logic is one of the best known and most studied of the approaches to nonmonotonic reasoning. Several variants of default logic have subsequently been proposed to give systems with properties differing from the original. In this paper, we examine the relationship between default logic and its major variants. We accomplish this by translating a default theory under a variant interpretation into a second default theory, under the original Reiter semantics, wherein the variant interpretation is respected. That is, in each case we show that, given an extension of a translated theory, one may extract an extension of the original variant default logic theory. We show how constrained, rational, justified, and cumulative default logic can be expressed in Reiter's default logic. As well, we show how Reiter's default logic can be expressed in rational default logic. From this, we suggest that any such variant can be similarly treated. Consequently, we provide a unification of default logics, showing how the original formulation of default logic may express its variants. Moreover, the translations clearly express the relationships between alternative approaches to default logic. The translations themselves are shown to generally have good properties. Thus, in at least a theoretical sense, we show that these variants are in a sense superfluous, in that for any of these variants of default logic, we can exactly mimic the behaviour of a variant in standard default logic. As well, the translations lend insight into means of classifying the expressive power of default logic variants; specifically we suggest that the property of semi-monotonicity represents a division with respect to expressibility, whereas regularity and cumulativity do not.

Unsupervised outlier detection using random subspace and subsampling ensembles of Dirichlet process mixtures

With the widespread use of closed-circuit television (CCTV) surveillance systems in public areas, crowd anomaly detection has become an increasingly critical aspect of the intelligent video surveillance system. It requires workforce and continuous attention to decide on the captured event, which is hard to perform by individuals. The available literature on human action detection includes various approaches to detect abnormal crowd behavior, which is articulated as an outlier detection problem. This paper presents a detailed review of the recent development of anomaly detection methods from the perspectives of computer vision on different available datasets. A new taxonomic organization of existing works in crowd analysis and anomaly detection has been introduced. A summarization of existing reviews and datasets related to anomaly detection has been listed. It covers an overview of different crowd concepts, including mass gathering events analysis and challenges, types of anomalies, and surveillance systems. Additionally, research trends and future work prospects have been analyzed.

The outlier detection problem has important applications in the field of fraud detection, network robustness analysis, and intrusion detection. Most such applications are high dimensional domains in which the data can contain hundreds of dimensions. Many recent algorithms use concepts of proximity in order to find outliers based on their relationship to the rest of the data. However, in high dimensional space, the data is sparse and the notion of proximity fails to retain its meaningfulness. In fact, the sparsity of high dimensional data implies that every point is an almost equally good outlier from the perspective of proximity-based definitions. Consequently, for high dimensional data, the notion of finding meaningful outliers becomes substantially more complex and non-obvious. In this paper, we discuss new techniques for outlier detection which find the outliers by studying the behavior of projections from the data set.

Over the last few years, Kernel Principal Component Analysis (KPCA) has found several applications in outlier detection. A relatively recent method uses KPCA to compute the reconstruction error (RE) of previously unseen samples and, via thresholding, to identify atypical samples. In this paper we propose an alternative method, which performs the same task, but considers Mahalanobis distances in the orthogonal complement of the subspace that is utilized to compute the reconstruction error. In order to illustrate its merits, we provide qualitative and quantitative results on both artificial and real datasets and we show that it is competitive, if not superior, for several outlier detection tasks, when compared to the original RE-based variant and the One-Class SVM detection approach.

Converting SVDD scores into probability estimates: Application to outlier detection

The paper presents the problem of outlier detection in the rule knowledge bases. Unusual (rare) rules, regarded here as the deviation, should be the subject of analysis experts and knowledge engineers because they can influence the efficiency of inference in decision support systems. The author presents a different approach in finding outliers in the rules. The experiments with their results are also presented in the paper.

This paper is concerned with the detection of dim targets in cluttered image sequences. It is an extension of our previous work [7] in which we viewed target detection as an outlier detection problem. In that work the background was modelled by a uni-modal Gaussian. In this paper a Gaussian mixture-model is used to describe the background in which the the number of components is automatically selected. As an outlier does not automatically imply a target, a final stage has been added in which all points below a set density function value are passed to a support vector classifier to be identified as a target or background. This system is compared favourably to a baseline technique [12].

The classification performances of the traditional one-class support vector machine (OCSVM) and its variants are often not satisfying when outliers are complex. To deal with this case, assigning smaller weights to these outliers may alleviate their influence upon the classification boundary and enhance the robustness of OCSVM. In this paper, a novel adaptive-weighted one-class support vector machine (AWOCSVM) is proposed for dealing with the outlier detection problem. The appropriate weights are assigned to training samples by considering both their local densities and distances between them to their center. Experimental results on two synthetic data sets and eight benchmark data sets demonstrate that the proposed AWOCSVM achieves more compact classification boundary and superior performance in comparison with the traditional OCSVM and one related approach.

Parkinsonism as an outlier detection problem

Spatiotemporal data mining (STDM) discovers useful patterns from the dynamic interplay between space and time. Several available surveys capture STDM advances and report a wealth of important progress in this field. However, STDM challenges and problems are not thoroughly discussed and presented in articles of their own. We attempt to fill this gap by providing a comprehensive literature survey on state-of-the-art advances in STDM. We describe the challenging issues and their causes and open gaps of multiple STDM directions and aspects. Specifically, we investigate the challenging issues in regards to spatiotemporal relationships, interdisciplinarity, discretisation, and data characteristics. Moreover, we discuss the limitations in the literature and open research problems related to spatiotemporal data representations, modelling and visualisation, and comprehensiveness of approaches. We explain issues related to STDM tasks of classification, clustering, hotspot detection, association and pattern mining, outlier detection, visualisation, visual analytics, and computer vision tasks. We also highlight STDM issues related to multiple applications including crime and public safety, traffic and transportation, earth and environment monitoring, epidemiology, social media, and Internet of Things.

The problem of predicting image or video interestingness from their low-level feature representations has received increasing interest. As a highly subjective visual attribute, annotating the interestingness value of training data for learning a prediction model is challenging. To make the annotation less subjective and more reliable, recent studies employ crowdsourcing tools to collect pairwise comparisons – relying on majority voting to prune the annotation outliers/errors. In this paper, we propose a more principled way to identify annotation outliers by formulating the interestingness prediction task as a unified robust learning to rank problem, tackling both the outlier detection and interestingness prediction tasks jointly. Extensive experiments on both image and video interestingness benchmark datasets demonstrate that our new approach significantly outperforms state-of-the-art alternatives.

The problem of unsupervised outlier detection is challenging, especially when the structure of data is unknown. This paper presents a new density-based outlier detection technique that detects the top-n outliers. It overcomes the limitations of existing approaches, like low accuracy and high sensitivity to parameters. Our approach provides a score to each object called Dynamic-Window Outlier Factor (DWOF). DWOF improves Resolution-based Outlier Factor method (ROF) to consider varying-density clusters, which improves outliers’ ranking even when providing same outliers. Experiments show that DWOF’s average accuracy is better than existing approaches and less sensitive to its parameter.