Abstract When dealing with sensitive data in automated data-driven decision-making, an important concern is to learn predictors with high performance towards a class label, whilst minimising for the discrimination towards any sensitive attribute, like gender or race, induced from biased data. Hybrid tree optimisation criteria have been proposed which combine classification performance and fairness. Although the threshold-free ROC-AUC is the standard for measuring classification model performance, current fair tree classification methods mainly optimise for a fixed threshold on the fairness metric. In this paper, we propose SCAFF—splitting criterion AUC for Fairness—a compound decision tree splitting criterion which combines the threshold-free strong demographic parity with ROC-AUC termed, easily applicable as an ensemble. Our method simultaneously leverages multiple sensitive attributes of which the values may be multicategorical, and is tunable with respect to the unavoidable performance-fairness trade-off. In our experiments, we demonstrate how SCAFF generates effective models with competitive performance and fairness with respect to binary, multicategorical, and multiple sensitive attributes.

As of 2021, more demanding NOx emission requirements entered into force for newly built ships operating on the North and Baltic Sea. Even though various methods are used to assess ships' pollution in ports and off the coastal areas, monitoring over the open sea has been infeasible until now. In this work, we present a novel automated method for evaluation of NO2 emissions produced by individual seagoing ships. We use the spatial association statistic local Moran's I in order to improve the distinguishability between the plume and the background. Using the Automatic Identification Signal (AIS) data of ship locations as well as incorporated uncertainties in wind speed and wind direction, we automatically associate the detected plumes with individual ships. We evaluate the quality of ship-plume matching by calculating the Pearson correlation coefficient between the values of a model-based emission proxy and the estimated NO2 concentrations. For five of the six analyzed areas, our method yields results that are an improvement over the baseline approach used in a previous study.

Transit of wasteful materials within the European Union is highly regulated through a system of permits. Waste processing costs vary greatly depending on the waste category of a permit. Therefore, companies may have a financial incentive to allege transporting waste with erroneous categorisation. Our goal is to assist inspectors in selecting potentially manipulated permits for further investigation, making their task more effective and efficient. Due to data limitations, a supervised learning approach based on historical cases is not possible. Standard unsupervised approaches, such as outlier detection and data quality-assurance techniques, are not suited since we are interested in targeting non-random modifications in both category and category-correlated features. For this purpose we (1) introduce the concept of crosslier: an anomalous instance of a category which lies across other categories; (2) propose eXPose: a novel approach to crosslier detection based on supervised category modelling; and (3) present the crosslier diagram: a visualisation tool specifically designed for domain experts to easily assess crossliers. We compare eXPose against traditional outlier detection methods in various benchmark datasets with synthetic crossliers and show the superior performance of our method in targeting these instances.

This paper studies automatic image classification by modeling soft assignment in the popular codebook model. The codebook model describes an image as a bag of discrete visual words selected from a vocabulary, where the frequency distributions of visual words in an image allow classification. One inherent component of the codebook model is the assignment of discrete visual words to continuous image features. Despite the clear mismatch of this hard assignment with the nature of continuous features, the approach has been successfully applied for some years. In this paper, we investigate four types of soft assignment of visual words to image features. We demonstrate that explicitly modeling visual word assignment ambiguity improves classification performance compared to the hard assignment of the traditional codebook model. The traditional codebook model is compared against our method for five well-known data sets: 15 natural scenes, Caltech-101, Caltech-256, and Pascal VOC 2007/2008. We demonstrate that large codebook vocabulary sizes completely deteriorate the performance of the traditional model, whereas the proposed model performs consistently. Moreover, we show that our method profits in high-dimensional feature spaces and reaps higher benefits when increasing the number of image categories.

Whereas video tells a narrative by a composition of shots, current video retrieval methods focus mainly on single shots. In retrieval performance estimation, similar shots in a narrative may result in performance overestimation. We propose an episode-based version of cross-validation leading up to 14% classification improvement over shot-based cross-validation.

File carving is the process of recovering files from a disk without the help of a file system. In forensics, it is a helpful tool in finding hidden or recently removed disk content. Known signatures in file headers and footers are especially useful in carving such files out, that is, from header until footer. However, this approach assumes that file clusters remain in order. In case of file fragmentation, file clusters can be disconnected and the order can even be disrupted such that straighforward carving will fail. In this paper, we focus on methods for classifying clusters into file types by using the statistics of the clusters. By not exploiting the possible embedded signatures, we generate evidence from a different source that can be integrated later on. We propose a set of characteristic features and use statistical pattern recognition to learn a supervised classification model for a range of relevant file types. We exploit the statistics of a restricted number of neighboring clusters (context) to improve classification performance. In the experiments we show that the proposed features indeed enable the differentation of clusters into file types. Moreover, for some file types the incorporation of cluster context improves the recognition performance significantly.

Digital video is sequential in nature. When video data is used in a semantic concept classification task, the episodes are usually summarized with shots. The shots are annotated as containing, or not containing, a certain concept resulting in a labeled dataset. These labeled shots can subsequently be used by supervised learning methods (classifiers) where they are trained to predict the absence or presence of the concept in unseen shots and episodes. The performance of such automatic classification systems is usually estimated with cross-validation. By taking random samples from the dataset for training and testing as such, part of the shots from an episode are in the training set and another part from the same episode is in the test set. Accordingly, data dependence between training and test set is introduced, resulting in too optimistic performance estimates. In this paper, we experimentally show this bias, and propose how this bias can be prevented using episode-constrained crossvalidation. Moreover, we show that a 17% higher classifier performance can be achieved by using episode constrained cross-validation for classifier parameter tuning.

In this paper we focus on high dimensional data sets for which the number of dimensions is an order of magnitude higher than the number of objects. From a classifier design standpoint, such small sample size problems have some interesting challenges. First, in any subspace with as many dimensions as objects the data set can be separated with an almost arbitrary linear hyperplane. Second, another important issue is to determine which features are responsible for the phenomenon under consideration. This problem comes down to finding as few features as possible that still can discriminate the classes involved. To attack these problems, we propose the LESS (lowest error in a sparse subspace) classifier. The LESS classifier is a weighted nearest mean classifier that efficiently finds linear discriminants in sparse subspaces, where the subspace is found automatically. In the experiments we compare LESS to related state-of-the-art classifiers like among others linear ridge regression with the LASSO and the support vector machine. It turns out that LESS performs competitively while it uses the fewest features.

MOTIVATION: Microarray gene expression data are increasingly employed to identify sets of marker genes that accurately predict disease development and outcome in cancer. Many computational approaches have been proposed to construct such predictors. However, there is, as yet, no objective way to evaluate whether a new approach truly improves on the current state of the art. In addition no 'standard' computational approach has emerged which enables robust outcome prediction. RESULTS: An important contribution of this work is the description of a principled training and validation protocol, which allows objective evaluation of the complete methodology for constructing a predictor. We review the possible choices of computational approaches, with specific emphasis on predictor choice and reporter selection strategies. Employing this training-validation protocol, we evaluated different reporter selection strategies and predictors on six gene expression datasets of varying degrees of difficulty. We demonstrate that simple reporter selection strategies (forward filtering and shrunken centroids) work surprisingly well and outperform partial least squares in four of the six datasets. Similarly, simple predictors, such as the nearest mean classifier, outperform more complex classifiers. Our training-validation protocol provides a robust methodology to evaluate the performance of new computational approaches and to objectively compare outcome predictions on different datasets.

We present the Nearest Subclass Classifier (NSC), which is a classification algorithm that unifies the flexibility of the nearest neighbor classifier with the robustness of the nearest mean classifier. The algorithm is based on the Maximum Variance Cluster algorithm and, as such, it belongs to the class of prototype-based classifiers. The variance constraint parameter of the cluster algorithm serves to regularize the classifier, that is, to prevent overfitting. With a low variance constraint value, the classifier turns into the nearest neighbor classifier and, with a high variance parameter, it becomes the nearest mean classifier with the respective properties. In other words, the number of prototypes ranges from the whole training set to only one per class. In the experiments, we compared the NSC with regard to its performance and data set compression ratio to several other prototype-based methods. On several data sets, the NSC performed similarly to the k-nearest neighbor classifier, which is a well-established classifier in many domains. Also concerning storage requirements and classification speed, the NSC has favorable properties, so it gives a good compromise between classification performance and efficiency.

In this paper, we specifically focus on high-dimensional data sets for which the number of dimensions is an order of magnitude higher than the number of objects. From a classifier design standpoint, such small sample size problems have some interesting challenges. The first challenge is to find, from all hyperplanes that separate the classes, a separating hyperplane which generalizes well for future data. A second important task is to determine which features are required to distinguish the classes. To attack these problems, we propose the LESS (Lowest Error in a Sparse Subspace) classifier that efficiently finds linear discriminants in a sparse subspace. In contrast with most classifiers for high-dimensional data sets, the LESS classifier incorporates a (simple) data model. Further, by means of a regularization parameter, the classifier establishes a suitable trade-off between subspace sparseness and classification accuracy. In the experiments, we show how LESS performs on several high-dimensional data sets and compare its performance to related state-of-the-art classifiers like, among others, linear ridge regression with the LASSO and the Support Vector Machine. It turns out that LESS performs competitively while using fewer dimensions.

Clustering is inherently a difficult problem, both with respect to the definition of adequate models as well as to the optimization of the models. We present a model for the cluster problem that does not need knowledge about the number of clusters a priori. This property is among others useful in the image segmentation domain, which we especially address. Further, we propose a cellular coevolutionary algorithm for the optimization of the model. Within this scheme multiple agents are placed in a regular two-dimensional (2-D) grid representing the image, which imposes neighboring relations on them. The agents cooperatively consider pixel migration from one agent to the other in order to improve the homogeneity of the ensemble of the image regions they represent. If the union of the regions of neighboring agents is homogeneous then the agents form alliances. On the other hand, if an agent discovers a deviant subject, it isolates the subject. In the experiments we show the effectiveness of the proposed method and compare it to other segmentation algorithms. The efficiency can easily be improved by exploiting the intrinsic parallelism of the proposed method.

We present a partitional cluster algorithm that minimizes the sum-of-squared-error criterion while imposing a hard constraint on the cluster variance. Conceptually, hypothesized clusters act in parallel and cooperate with their neighboring clusters in order to minimize the criterion and to satisfy the variance constraint. In order to enable the demarcation of the cluster neighborhood without crucial parameters, we introduce the notion of foreign cluster samples. Finally, we demonstrate a new method for cluster tendency assessment based on varying the variance constraint parameter.

Studies the motion correspondence problem for which a diversity of qualitative and statistical solutions exist. We concentrate on qualitative modeling, especially in situations where assignment conflicts arise either because multiple features compete for one detected point or because multiple detected points fit a single feature point. We leave out the possibility of point track initiation and termination because that principally conflicts with allowing for temporary point occlusion. We introduce individual, combined, and global motion models and fit existing qualitative solutions in this framework. Additionally, we present a tracking algorithm that satisfies these-possibly constrained-models in a greedy matching sense, including an effective way to handle detection errors and occlusion. The performance evaluation shows that the proposed algorithm outperforms existing greedy matching algorithms. Finally, we describe an extension to the tracker that enables automatic initialization of the point tracks. Several experiments show that the extended algorithm is efficient, hardly sensitive to its few parameters, and qualitatively better than other algorithms, including the presumed optimal statistical multiple hypothesis tracker.