Search

In this paper, we propose a novel approach for outlier detection, called local projections, which is based on concepts of Local Outlier Factor (LOF) (Breunig et al., 2000) and RobPCA (Hubert et al., 2005). By using aspects of both methods, our algorithm is robust towards noise variables and is capable of performing outlier detection in multi-group situations. We are further not reliant on a specific underlying data distribution. For each observation of a dataset, we identify a local group of dense nearby observations, which we call a core, based on a modification of the k-nearest neighbours algorithm. By projecting the dataset onto the space spanned by those observations, two aspects are revealed. First, we can analyze the distance from an observation to the center of the core within the projection space in order to provide a measure of quality of description of the observation by the projection. Second, we consider the distance of the observation to the projection space in order to assess the suitability of the core for describing the outlyingness of the observation. These novel interpretations lead to a univariate measure of outlyingness based on aggregations over all local projections, which outperforms LOF and RobPCA as well as other popular methods like PCOut (Filzmoser et al., 2008) and subspace-based outlier detection (Kriegel et al., 2009) in our simulation setups. Experiments in the context of real-word applications employing datasets of various dimensionality demonstrate the advantages of local projections.

A powerful data transformation method named guided projections is proposed creating new possibilities to reveal the group structure of high-dimensional data in the presence of noise variables. Using projections onto a space spanned by a selection of a small number of observations allows measuring the similarity of other observations to the selection based on orthogonal and score distances. Observations are iteratively exchanged from the selection creating a nonrandom sequence of projections, which we call guided projections. In contrast to conventional projection pursuit methods, which typically identify a low-dimensional projection revealing some interesting features contained in the data, guided projections generate a series of projections that serve as a basis not just for diagnostic plots but to directly investigate the group structure in data. Based on simulated data, we identify the strengths and limitations of guided projections in comparison to commonly employed data transformation methods...

Media content in large repositories usually exhibits multiple groups of strongly varying sizes. Media of potential interest often form notably smaller groups. Such media groups differ so much from the remaining data that it may be worthy to look at them in more detail. In contrast, media with popular content appear in larger groups. Identifying groups of varying sizes is addressed by clustering of imbalanced data. Clustering highly imbalanced media groups is additionally challenged by the high dimensionality of the underlying features. In this paper, we present the imbalanced clustering (IClust) algorithm designed to reveal group structures in high-dimensional media data. IClust employs an existing clustering method in order to find an initial set of a large number of potentially highly pure clusters which are then successively merged. The main advantage of IClust is that the number of clusters does not have to be pre-specified and that no specific assumptions about the cluster or data characteristics need to be made. Experiments on real-world media data demonstrate that in comparison to existing methods, IClust is able to better identify media groups, especially groups of small sizes.

We introduce a robust k-means-based clustering method for high-dimensional data where not only outliers but also a large number of noise variables are very likely to be present [4]. Although Kondo et al. [2] already addressed such an application scenario, our approach goes even further. Firstly, the introduced method is designed to identify clusters, informative variables, and outliers simultaneously. Secondly, the proposed clustering technique additionally aims at optimizing required parameters, e.g. the number of clusters. This is a great advantage over most existing methods. Moreover, the robustness aspect is achieved through a robust initialization [3] and a proposed weighting function using the Local Outlier Factor [1]. The weighting function provides a valuable source of information about the outlyingness of each observation for a subsequent outlier detection. In order to reveal both clusters and informative variables properly, the approach uses a lasso-type penalty [5]. The method has thoroughly been tested on simulated as well as on real high-dimensional datasets. The conducted experiments demonstrated a great ability of the clustering method to identify clusters, outliers, and informative variables.

Panoramic camera systems on robots exploring the surface of Mars are used to collect images of terrain and rock outcrops which they encounter along their traverse. Image mosaics from these cameras are essential in mapping the surface geology and selecting locations for analysis by other instruments on the rover's payload. 2‐D images do not truly portray the depth of field of features within an image, nor their 3‐D geometry. This paper describes a new 3‐D visualization software tool for geological analysis of Martian rover‐derived Digital Outcrop Models created using photogrammetric processing of stereo‐images using the Planetary Robotics Vision Processing tool developed for 3‐D vision processing of ExoMars PanCam and Mars 2020 Mastcam‐Z data. Digital Outcrop Models are rendered in real time in the Planetary Robotics 3‐D Viewer PRo3D, allowing scientists to roam outcrops as in a terrestrial field campaign. Digitization of point, line, and polyline features is used for measuring the physical dimensions of geological features and communicating interpretations. Dip and strike of bedding and fractures is measured by digitizing a polyline along the contact or fracture trace, through which a best fit plane is plotted. The attitude of this plane is calculated in the software. Here we apply these tools to analysis of sedimentary rock outcrops and quantification of the geometry of fracture systems encountered by the science teams of NASA's Mars Exploration Rover Opportunity and Mars Science Laboratory rover Curiosity. We show the benefits PRo3D allows for visualization and collection of geological interpretations and analyses from rover‐derived stereo‐images., Earth and Space Science, 5 (7), ISSN:2333-5084

Testing for structural changes is a well studied field. Classical tests for breakpoint detection utilize F-statistics which depend on independent and identically normal distributed residuals. In general, this condition is not satisfied which leads to distorted test results when the p-values of classical tests are close to the significance level. Thus, permutation tests are used to properly estimate the critical values. Based on the accounting data of Austrian hospitals, collected by the Austrian social insurance institutions, specific observations (hospital stays) which are connected to pre-defined diseases are analysed. For those groups of observations we use characteristic factors, to test for structural changes from different perspectives. The first test analyses the temporal trend and identifies breakpoints, caused by changes in the underlying system between years. The second analysis focuses on identifying differences between hospitals. Both implemented tests ensure the often ignored aspect of a homogeneous data base for further analysis.

Percutaneous retrograde screw fixation for acetabular fractures is a demanding procedure due to the complex anatomy of the pelvis and the varying narrow safe bony corridors. Limited information is available on optimal screw placement and the geometry of safe zones for screw insertion in the pelvis. Three-dimensional reconstructions of 50 consecutive CT scans of polytrauma patients (35 males, 15 females) were used to introduce three virtual CAD bolts (representing screws) into the anterior column (superior ramus of the pubic bone), posterior column (the ischial bone) and the supraacetabular region, as performed during percutaneous screw fixation. The three-dimensional (3D) position of these screws was evaluated with a computer software (MIMICS) after virtual optimal insertion. The 3D position, the narrowest zone and the distance to the hip joint of the two columns and the supraacetabular region were defined. The mean maximal screw length for the three virtual screws measured between 107.4 and 148 ± 18.7 mm. The narrowest zone of the pelvic bone (superior pubic ramus) had a width of 9.2 ± 2.4 mm. The average distances between the bolts and the hip joint were 3.9 and 19.4 ± 7.4 mm. For the anterior column (superior pubic ramus) screw, the mean lateral angle to the sagittal midline plane was 39.0 ± 3.2° and the mean posterior angle to the transversal midline plane was 15.1 ± 4.0°. The mean supraacetabular screw angles measured 22.4 ± 3.4° (medial), 35.3 ± 4.6° (cranial) and the mean angles for the ischial screw were 12.0 ± 5.4° (posterior) and 18.4 ± 4.0° (lateral). The zones for safe screw positioning are very narrow, making percutaneous screw fixation of the acetabulum a challenging procedure. The predefined angles for the most frequently positioned percutaneous screws may aid in preoperative planning, decrease operative and radiation times and help to increase safe insertion of screws.

3D visibility analysis plays a key role in urban planning for assessing the visual impact of proposed buildings on the cityscape. A call for proposals typically yields around 30 candidate buildings that need to be evaluated with respect to selected viewpoints. Current visibility analysis methods are very time-consuming and limited to a small number of viewpoints. Further, analysts neither have measures to evaluate candidates quantitatively, nor to compare them efficiently. The primary contribution of this work is the design study of Vis-A-Ware, a visualization system to qualitatively and quantitatively evaluate, rank, and compare visibility data of candidate buildings with respect to a large number of viewpoints. Vis-A-Ware features a 3D spatial view of an urban scene and non-spatial views of data derived from visibility evaluations, which are tightly integrated by linked interaction. To enable a quantitative evaluation we developed four metrics in accordance with experts from urban planning. We illustrate the applicability of Vis-A-Ware on the basis of a use case scenario and present results from informal feedback sessions with domain experts from urban planning and development. This feedback suggests that Vis-A-Ware is a valuable tool for visibility analysis allowing analysts to answer complex questions more efficiently and objectively.

Background FTY720 (Fingolimod) is a novel immunosuppressive drug investigated in clinical trials for organ transplantation and multiple sclerosis. It acts as a functional sphingosine-1-phosphate (S1P) receptor antagonist, thereby inhibiting the egress of lymphocytes from secondary lymphoid organs. As S1P is able to prevent IL-1beta induced cartilage degradation, we examined the direct impact of FTY720 on cytokine induced cartilage destruction. Methods Bovine chondrocytes were treated with the bioactive phosphorylated form of FTY720 (FTY720-P) in combination with IL-1beta or TNF-alpha. Expression of MMP-1,-3.-13, iNOS and ADAMTS-4,-5 and COX-2 was evaluated using quantitative real-time PCR and western blot. Glycosaminoglycan depletion from cartilage explants was determined using a 1,9-dimethylene blue assay and safranin O staining. Results FTY720-P significantly reduced IL-1beta and TNF-alpha induced expression of iNOS. In contrast FTY720-P increased MMP-3 and ADAMTS-5 mRNA expression. Furthermore depletion of glycosaminoglycan from cartilage explants by IL-1beta and TNF-alpha was significantly enhanced by FTY720-P in an MMP-3 dependent manner. Conclusions Our results suggest that FTY720 may enhance cartilage degradation in pro-inflammatory environment.