@article{ZhelayskayaVasileShpritsetal.2019, author = {Zhelayskaya, Irina S. and Vasile, Ruggero and Shprits, Yuri Y. and Stolle, Claudia and Matzka, J{\"u}rgen}, title = {Systematic Analysis of Machine Learning and Feature Selection Techniques for Prediction of the Kp Index}, series = {Space Weather: The International Journal of Research and Applications}, volume = {17}, journal = {Space Weather: The International Journal of Research and Applications}, number = {10}, publisher = {American Geophysical Union}, address = {Washington}, issn = {1542-7390}, doi = {10.1029/2019SW002271}, pages = {1461 -- 1486}, year = {2019}, abstract = {The Kp index is a measure of the midlatitude global geomagnetic activity and represents short-term magnetic variations driven by solar wind plasma and interplanetary magnetic field. The Kp index is one of the most widely used indicators for space weather alerts and serves as input to various models, such as for the thermosphere and the radiation belts. It is therefore crucial to predict the Kp index accurately. Previous work in this area has mostly employed artificial neural networks to nowcast Kp, based their inferences on the recent history of Kp and on solar wind measurements at L1. In this study, we systematically test how different machine learning techniques perform on the task of nowcasting and forecasting Kp for prediction horizons of up to 12 hr. Additionally, we investigate different methods of machine learning and information theory for selecting the optimal inputs to a predictive model. We illustrate how these methods can be applied to select the most important inputs to a predictive model of Kp and to significantly reduce input dimensionality. We compare our best performing models based on a reduced set of optimal inputs with the existing models of Kp, using different test intervals, and show how this selection can affect model performance.}, language = {en} } @phdthesis{Zhelavskaya2020, author = {Zhelavskaya, Irina}, title = {Modeling of the Plasmasphere Dynamics}, doi = {10.25932/publishup-48243}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-482433}, school = {Universit{\"a}t Potsdam}, pages = {xlii, 256}, year = {2020}, abstract = {The plasmasphere is a dynamic region of cold, dense plasma surrounding the Earth. Its shape and size are highly susceptible to variations in solar and geomagnetic conditions. Having an accurate model of plasma density in the plasmasphere is important for GNSS navigation and for predicting hazardous effects of radiation in space on spacecraft. The distribution of cold plasma and its dynamic dependence on solar wind and geomagnetic conditions remain, however, poorly quantified. Existing empirical models of plasma density tend to be oversimplified as they are based on statistical averages over static parameters. Understanding the global dynamics of the plasmasphere using observations from space remains a challenge, as existing density measurements are sparse and limited to locations where satellites can provide in-situ observations. In this dissertation, we demonstrate how such sparse electron density measurements can be used to reconstruct the global electron density distribution in the plasmasphere and capture its dynamic dependence on solar wind and geomagnetic conditions. First, we develop an automated algorithm to determine the electron density from in-situ measurements of the electric field on the Van Allen Probes spacecraft. In particular, we design a neural network to infer the upper hybrid resonance frequency from the dynamic spectrograms obtained with the Electric and Magnetic Field Instrument Suite and Integrated Science (EMFISIS) instrumentation suite, which is then used to calculate the electron number density. The developed Neural-network-based Upper hybrid Resonance Determination (NURD) algorithm is applied to more than four years of EMFISIS measurements to produce the publicly available electron density data set. We utilize the obtained electron density data set to develop a new global model of plasma density by employing a neural network-based modeling approach. In addition to the location, the model takes the time history of geomagnetic indices and location as inputs, and produces electron density in the equatorial plane as an output. It is extensively validated using in-situ density measurements from the Van Allen Probes mission, and also by comparing the predicted global evolution of the plasmasphere with the global IMAGE EUV images of He+ distribution. The model successfully reproduces erosion of the plasmasphere on the night side as well as plume formation and evolution, and agrees well with data. The performance of neural networks strongly depends on the availability of training data, which is limited during intervals of high geomagnetic activity. In order to provide reliable density predictions during such intervals, we can employ physics-based modeling. We develop a new approach for optimally combining the neural network- and physics-based models of the plasmasphere by means of data assimilation. The developed approach utilizes advantages of both neural network- and physics-based modeling and produces reliable global plasma density reconstructions for quiet, disturbed, and extreme geomagnetic conditions. Finally, we extend the developed machine learning-based tools and apply them to another important problem in the field of space weather, the prediction of the geomagnetic index Kp. The Kp index is one of the most widely used indicators for space weather alerts and serves as input to various models, such as for the thermosphere, the radiation belts and the plasmasphere. It is therefore crucial to predict the Kp index accurately. Previous work in this area has mostly employed artificial neural networks to nowcast and make short-term predictions of Kp, basing their inferences on the recent history of Kp and solar wind measurements at L1. We analyze how the performance of neural networks compares to other machine learning algorithms for nowcasting and forecasting Kp for up to 12 hours ahead. Additionally, we investigate several machine learning and information theory methods for selecting the optimal inputs to a predictive model of Kp. The developed tools for feature selection can also be applied to other problems in space physics in order to reduce the input dimensionality and identify the most important drivers. Research outlined in this dissertation clearly demonstrates that machine learning tools can be used to develop empirical models from sparse data and also can be used to understand the underlying physical processes. Combining machine learning, physics-based modeling and data assimilation allows us to develop novel methods benefiting from these different approaches.}, language = {en} } @article{RauschBrockmeyerSchwerdtle2021, author = {Rausch, Ann-Kristin and Brockmeyer, Robert and Schwerdtle, Tanja}, title = {Development and validation of a liquid chromatography tandem mass spectrometry multi-method for the determination of 41 free and modified mycotoxins in beer}, series = {Food chemistry}, volume = {338}, journal = {Food chemistry}, publisher = {Elsevier}, address = {New York, NY}, issn = {1873-7072}, doi = {10.1016/j.foodchem.2020.127801}, year = {2021}, abstract = {A fast high performance liquid chromatography tandem mass spectrometry multi-method based on an ACN-precipitation extraction was developed for the analysis of 41 (modified) mycotoxins in beer. Validation according to the performance criteria defined by the European Commission (EC) in Commission Decision no. 657/2002 revealed good linearity (R2 > 0.99), repeatability (RSDr < 15\%), reproducibility (RSDR < 15\%), and recovery (79-100\%). Limits of quantification ranging from 0.04 to 75 µg/L were obtained. Matrix effects varied from -67 to +319\% and were compensated for using standard addition. In total, 87 beer samples, produced worldwide, were analyzed for the presence of mycotoxins with a focus on modified mycotoxins, whereof 76\% of the samples were contaminated with at least one mycotoxin. The most prevalent mycotoxins were deoxynivalenol-3-glucoside (63\%), HT-2 toxin (15\%), and tenuazonic acid (13\%). Exposure estimates of deoxynivalenol and its metabolites for German beer revealed no significant contribution to intake of deoxynivalenol.}, language = {en} } @article{PlikkEngelsLuotoetal.2019, author = {Plikk, Anna and Engels, Stefan and Luoto, Tomi P. and Nazarova, Larisa B. and Salonen, J. Sakari and Helmens, Karin F.}, title = {Chironomid-based temperature reconstruction for the Eemian Interglacial (MIS 5e) at Sokli, northeast Finland}, series = {Journal of paleolimnology}, volume = {61}, journal = {Journal of paleolimnology}, number = {3}, publisher = {Springer Science}, address = {Dordrecht}, issn = {0921-2728}, doi = {10.1007/s10933-018-00064-y}, pages = {355 -- 371}, year = {2019}, abstract = {The Last Interglacial (Eemian, MIS 5e) can be considered a test-bed for climate dynamics under a warmer-than-present climate. In this study we present a chironomid record from the high latitude Sokli site (N Finland), where a long continuous sediment sequence from the last interglacial has been preserved from glacial erosion. The chironomid-analysis shows a diverse fauna, with dominance of warm-water indicators and shifts in assemblage composition that can be attributed to temperature, lake depth, productivity and habitat availability. Quantitative mean July paleotemperature estimates based on the chironomid data indicate overall mean July air temperatures up to 1 degrees C warmer than present. Two cooling events can be discerned, the Tunturi event, dated to about 127.5kaBP, in the lower part of the sequence, and the Varrio event, dated to about 119kaBP, associated with the beginning of a cooling trend in the upper part of the record. Warm conditions already at the onset of the interglacial contrast with a recent chironomid-based last interglacial temperature reconstruction from Denmark, which suggests a late onset of Eemian warming. The relatively small increase in inferred temperatures compared to present day temperatures at Sokli differs from other high latitude Eemian sites, and likely reflects the influence of the Atlantic Meridional Overturning Circulation in maintaining already elevated temperatures in Fennoscandia during interglacials.}, language = {en} } @article{KuninaHabenichtHautzKniggeetal.2015, author = {Kunina-Habenicht, Olga and Hautz, Wolf E. and Knigge, Michel and Spies, Claudia and Ahlers, Olaf}, title = {Assessing clinical reasoning (ASCLIRE): Instrument development and validation}, series = {Advances in health sciences education : theory and practice}, volume = {20}, journal = {Advances in health sciences education : theory and practice}, number = {5}, publisher = {Springer}, address = {Dordrecht}, issn = {1382-4996}, doi = {10.1007/s10459-015-9596-y}, pages = {1205 -- 1224}, year = {2015}, abstract = {Clinical reasoning is an essential competency in medical education. This study aimed at developing and validating a test to assess diagnostic accuracy, collected information, and diagnostic decision time in clinical reasoning. A norm-referenced computer-based test for the assessment of clinical reasoning (ASCLIRE) was developed, integrating the entire clinical decision process. In a cross-sectional study participants were asked to choose as many diagnostic measures as they deemed necessary to diagnose the underlying disease of six different cases with acute or sub-acute dyspnea and provide a diagnosis. 283 students and 20 content experts participated. In addition to diagnostic accuracy, respective decision time and number of used relevant diagnostic measures were documented as distinct performance indicators. The empirical structure of the test was investigated using a structural equation modeling approach. Experts showed higher accuracy rates and lower decision times than students. In a cross-sectional comparison, the diagnostic accuracy of students improved with the year of study. Wrong diagnoses provided by our sample were comparable to wrong diagnoses in practice. We found an excellent fit for a model with three latent factors-diagnostic accuracy, decision time, and choice of relevant diagnostic information-with diagnostic accuracy showing no significant correlation with decision time. ASCLIRE considers decision time as an important performance indicator beneath diagnostic accuracy and provides evidence that clinical reasoning is a complex ability comprising diagnostic accuracy, decision time, and choice of relevant diagnostic information as three partly correlated but still distinct aspects.}, language = {en} } @misc{HaynBeirleHamprechtetal.2009, author = {Hayn, Michael and Beirle, Steffen and Hamprecht, Fred A. and Platt, Ulrich and Menze, Bj{\"o}rn H. and Wagner, Thomas}, title = {Analysing spatio-temporal patterns of the global NO2-distribution retrieved from GOME satellite observations using a generalized additive model}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus-44999}, year = {2009}, abstract = {With the increasing availability of observational data from different sources at a global level, joint analysis of these data is becoming especially attractive. For such an analysis - oftentimes with little prior knowledge about local and global interactions between the different observational variables at hand - an exploratory, data-driven analysis of the data may be of particular relevance. In the present work we used generalized additive models (GAM) in an exemplary study of spatio-temporal patterns in the tropospheric NO2-distribution derived from GOME satellite observations (1996 to 2001) at global scale. We focused on identifying correlations between NO2 and local wind fields, a quantity which is of particular interest in the analysis of spatio-temporal interactions. Formulating general functional, parametric relationships between the observed NO2 distribution and local wind fields, however, is difficult - if not impossible. So, rather than following a modelbased analysis testing the data for predefined hypotheses (assuming, for example, sinusoidal seasonal trends), we used a GAM with non-parametric model terms to learn this functional relationship between NO2 and wind directly from the data. The NO2 observations showed to be affected by winddominated processes over large areas. We estimated the extent of areas affected by specific NO2 emission sources, and were able to highlight likely atmospheric transport "pathways". General temporal trends which were also part of our model - weekly, seasonal and linear changes - showed to be in good agreement with previous studies and alternative ways of analysing the time series. Overall, using a non-parametric model provided favorable means for a rapid inspection of this large spatio-temporal NO2 data set, with less bias than parametric approaches, and allowing to visualize dynamical processes of the NO2 distribution at a global scale.}, language = {en} } @article{FreitasdaCruzPfahringerMartensenetal.2021, author = {Freitas da Cruz, Harry and Pfahringer, Boris and Martensen, Tom and Schneider, Frederic and Meyer, Alexander and B{\"o}ttinger, Erwin and Schapranow, Matthieu-Patrick}, title = {Using interpretability approaches to update "black-box" clinical prediction models}, series = {Artificial intelligence in medicine : AIM}, volume = {111}, journal = {Artificial intelligence in medicine : AIM}, publisher = {Elsevier}, address = {Amsterdam}, issn = {0933-3657}, doi = {10.1016/j.artmed.2020.101982}, pages = {13}, year = {2021}, abstract = {Despite advances in machine learning-based clinical prediction models, only few of such models are actually deployed in clinical contexts. Among other reasons, this is due to a lack of validation studies. In this paper, we present and discuss the validation results of a machine learning model for the prediction of acute kidney injury in cardiac surgery patients initially developed on the MIMIC-III dataset when applied to an external cohort of an American research hospital. To help account for the performance differences observed, we utilized interpretability methods based on feature importance, which allowed experts to scrutinize model behavior both at the global and local level, making it possible to gain further insights into why it did not behave as expected on the validation cohort. The knowledge gleaned upon derivation can be potentially useful to assist model update during validation for more generalizable and simpler models. We argue that interpretability methods should be considered by practitioners as a further tool to help explain performance differences and inform model update in validation studies.}, language = {en} } @article{BehrendWarschburger2022, author = {Behrend, Nicole and Warschburger, Petra}, title = {Validation of a German version of the Body Appreciation Scale-2 (BAS-2)}, series = {Body image : an international journal of research}, volume = {41}, journal = {Body image : an international journal of research}, publisher = {Elsevier}, address = {Amsterdam}, issn = {1740-1445}, doi = {10.1016/j.bodyim.2022.01.020}, pages = {216 -- 224}, year = {2022}, abstract = {The Body Appreciation Scale-2 (BAS-2) is the most current measure of body appreciation, a central facet of positive body image. This work aimed to examine the factor structure and psychometric properties of a German version. In Study 1 (N = 659; M-age = 27.19, SD = 8.57), exploratory factor analyses (EFA) revealed that the German BAS-2 has a one-dimensional factor structure in women and men, showing cross-gender factor similarity. In Study 2 (N = 472; M-age = 30.08, SD = 12.35), confirmatory factor analysis (CFA) further supported the original scale's one-dimensional factor structure after freeing correlated errors. The German BAS-2 also showed partial scalar invariance across gender, with women and men not differing significantly in latent mean scores. As predicted, we found convergent relationships with measures of self-esteem, intuitive eating, and variables associated with negative body image (i.e., weight-and shape concerns, drive for thinness). Correlations with BMI were small and in an inverse direction. Incremental validity was demonstrated by predicting self-esteem and intuitive eating over and above measures of negative body image. Additionally, the German BAS-2 showed internal consistency and 2-week test-retest reliability. Overall, our results suggest that the German BAS-2 is a psychometrically sound instrument.}, language = {en} }