@misc{Brettraeger2025, type = {Master Thesis}, author = {Brettr{\"a}ger, Birgit}, title = {Untersuchung von Methoden zur Vorhersage der Ausleihzahlen von B{\"u}chern in {\"o}ffentlichen Bibliotheken anhand ihrer Coverbilder}, doi = {10.15771/MA_2025_4}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-20420}, school = {Technische Hochschule Wildau}, pages = {82}, year = {2025}, abstract = {In dieser Arbeit untersuche ich, ob anhand der Coverbilder von B{\"u}chern deren Ausleihh{\"a}ufigkeit in {\"o}ffentlichen Bibliotheken vorhergesagt werden kann. Zus{\"a}tzlich werte ich aus, ob ein Einbezug der Titel der B{\"u}cher sowie der Namen der Autor*innen die Vorhersagequalit{\"a}t verbessert. Dazu setze ich Werkzeuge aus dem Bereich Machine Learning ein. Mithilfe von Bild- und Textverarbeitungsverfahren (die Erstellung von Farbhistogrammen sowie die Verwendung von mit Deep Learning vortrainierten Modellen) erstelle ich Embeddings f{\"u}r die vorliegenden Daten. Die Ausleihh{\"a}ufigkeit versuche ich in Form einer Klassifikation vorherzusagen, und verwende daf{\"u}r einen KNN-Algorithmus. Die Vorhersagegenauigkeit (Accuracy) erreicht, unter Verwendung verschiedener Kombinationen von Embeddings und weiterer Parameter, nur einen H{\"o}chstwert von 65\%. Damit ist die Vorhersage zwar besser, als zuf{\"a}lliges Raten, aus meiner Sicht aber nicht ausreichend belastbar, um auf dieser Grundlage zum Beispiel Erwerbungsentscheidungen zu treffen.}, language = {de} } @inproceedings{FiebelkornKugelGuentheretal.2025, author = {Fiebelkorn, Richard and Kugel, Rafael and G{\"u}nther, Norman and Reiff-Stephan, J{\"o}rg}, title = {Predictive Maintenance in der Baumpflege - TreeAngel}, series = {Tagungsband AALE 2025 : menschenzentrierte Automation im digitalen Zeitalter}, booktitle = {Tagungsband AALE 2025 : menschenzentrierte Automation im digitalen Zeitalter}, editor = {Reiff-Stephan, J{\"o}rg and J{\"a}kel, Jens and Reichelt, Dirk}, publisher = {Hochschule f{\"u}r Technik, Wirtschaft und Kultur}, address = {Leipzig}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-20193}, pages = {251 -- 259}, year = {2025}, abstract = {Die Gew{\"a}hrleistung der Verkehrssicherheit von B{\"a}umen stellt eine zentrale Herausforderung f{\"u}r St{\"a}dte und Gemeinden dar. Traditionelle, manuelle Inspektionsmethoden sind zeit- und ressourcenintensiv sowie anf{\"a}llig f{\"u}r menschliche Fehler. Der folgende Artikel pr{\"a}sentiert ein innovatives System zur automatisierten Baumzustandsbewertung, welches moderne Kameratechnologien und K{\"u}nstliche Intelligenz (KI) nutzt. Im Rahmen einer Machbarkeitsstudie wurden Bilddaten, welche durch verschiedene Kamerasysteme erzeugt wurden, analysiert. Basierend auf diesen Daten wurde ein YOLOv8-Modell trainiert, welches eine pr{\"a}zise Detektion von B{\"a}umen und Sch{\"a}den, wie beispielsweise Totholz, erm{\"o}glicht. Die Ergebnisse des vorgestellten Prototypensystems sind vielversprechend hinsichtlich Genauigkeit und Effizienz. Es zeigt das Potenzial, manuelle Inspektionen durch automatisierte Verfahren zu erg{\"a}nzen oder zu ersetzen. Die pr{\"a}sentierten Resultate legen den Grundstein f{\"u}r nachhaltige und skalierbare Ans{\"a}tze in der Baumpflege und k{\"o}nnen zur Steigerung der {\"o}ffentlichen Sicherheit und Effizienz im urbanen Management beitragen.}, language = {de} } @article{NechyporenkoFrohmeStrelchuketal.2024, author = {Nechyporenko, Alina and Frohme, Marcus and Strelchuk, Yaroslav and Omelchenko, Vladyslav and Gargin, Vitaliy and Ishchenko, Liudmyla and Alekseeva, Victoriia}, title = {Galvanic Skin Response and Photoplethysmography for Stress Recognition Using Machine Learning and Wearable Sensors}, series = {Applied Sciences}, volume = {14}, journal = {Applied Sciences}, number = {24}, publisher = {MDPI}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-19803}, year = {2024}, abstract = {This study investigates stress recognition using galvanic skin response (GSR) and photoplethysmography (PPG) data and machine learning, with a new focus on air raid sirens as a stressor. It bridges laboratory and real-world conditions and highlights the reliability of wearable sensors in dynamic, high-stress environments such as war and conflict zones. The study involves 37 participants (20 men, 17 women), aged 20-30, who had not previously heard an air raid siren. A 70 dB "S-40 electric siren" (400-450 Hz) was delivered via headphones. The protocol included a 5 min resting period, followed by 3 min "no-stress" phase, followed by 3 min "stress" phase, and finally a 3 min recovery phase. GSR and PPG signals were recorded using Shimmer 3 GSR+ sensors on the fingers and earlobes. A single session was conducted to avoid sensitization. The workflow includes signal preprocessing to remove artifacts, feature extraction, feature selection, and application of different machine learning models to classify the "stress "and "no-stress" states. As a result, the best classification performance was shown by the k-Nearest Neighbors model, achieving 0.833 accuracy. This was achieved by using a particular combination of heart rate variability (HRV) and GSR features, which can be considered as new indicators of siren-induced stress.}, language = {en} } @inproceedings{ChumachenkoNechyporenkoFrohme2022, author = {Chumachenko, Dmytro and Nechyporenko, Alina and Frohme, Marcus}, title = {Impact of Russian War on COVID-19 Dynamics in Germany: the Simulation Study by Statistical Machine Learning}, series = {Proceedings of the 5th International Conference on Informatics \& Data-Driven Medicine, Lyon, France, November 18 - 20, 2022}, volume = {3302}, booktitle = {Proceedings of the 5th International Conference on Informatics \& Data-Driven Medicine, Lyon, France, November 18 - 20, 2022}, editor = {Shakhovska, Nataliya and Chr{\´e}tien, St{\´e}phane and Izonin, Ivan and Campos, Jaime}, publisher = {RWTH Aachen}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-19074}, pages = {78 -- 85}, year = {2022}, abstract = {The new coronavirus COVID-19 has been spreading worldwide for almost three years. The global community has developed effective measures to contain and control the pandemic. However, new factors are emerging that are driving the dynamics of COVID-19. One of these factors was the escalation of Russia's war in Ukraine. This study aims to test the hypothesis of the influence of migration flows caused by the Russian war in Ukraine on the dynamics of the epidemic process in Germany. For this, a model of the COVID-19 epidemic process was built based on the polynomial regression method. The model's adequacy was tested 30 days before the start of the escalation of the Russian war in Ukraine. To assess the impact of the war on the dynamics of COVID-19, the model was used to calculate the forecast of cumulative new and fatal cases of COVID-19 in Germany in the first 30 days after the start of the escalation of the Russian war in Ukraine. Modeling showed that migration flows from Ukraine are not a critical factor in the growth of the dynamics of the incidence of COVID-19 in Germany, but they influenced the number of cases. The next stage of the study is the development of more complex models for a detailed analysis of population dynamics, identifying factors influencing the epidemic process in the context of the Russian war in Ukraine, and assessing their information content.}, language = {en} } @misc{vandeSandReiffStephan2021, author = {van de Sand, Ron and Reiff-Stephan, J{\"o}rg}, title = {FrostByte Dataset}, doi = {10.15771/1894}, year = {2021}, abstract = {It is with great pleasure that we announce the release of the "Frost Byte" dataset. The dataset was collected during research on fault detection and diagnosis (FDD) approaches and their transferability to heterogeneous systems of industrial refrigeration systems at the Technical University of Applied Sciences Wildau. The data collection took place between 2018 and 2019 and is intended to supplement the ASHRAE 1043-RP dataset (Comstock \& Braun, 1999), which has been the only publicly available data source in this area to date. For comparison purposes, the method of data collection was chosen similarly, whereby this dataset collection was carried out using a ~ 100 kW refrigeration capacity ammonia system with plate heat exchangers. The dataset contains steady-state data equally sampled from five classes: 1. Normal (fault-free operating condition) 2. Reduced Condenser Water Flow 3. Reduced Evaporator Water Flow 4. Non-Condensable Gases 5. Refrigeration Leak Each class was investigated under changing operational conditions and varying fault severity levels (SL).}, language = {en} } @article{GavrylenkoPoltoratskyiNechyporenko2024, author = {Gavrylenko, Svitlana and Poltoratskyi, Vadym and Nechyporenko, Alina}, title = {Intrusion Detection Model Based on Improved Transformer}, series = {Advanced Information Systems}, volume = {8}, journal = {Advanced Information Systems}, number = {1}, publisher = {National Technical University Kharkiv Polytechnic Institute}, issn = {2522-9052}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-18750}, pages = {94 -- 99}, year = {2024}, abstract = {The object of the study is the process of identifying the state of a computer network. The subject of the study are the methods of identifying the state of computer networks. The purpose of the paper is to improve the efficacy of intrusion detection in computer networks by developing a method based on transformer models. The results obtained. The work analyzes traditional machine learning algorithms, deep learning methods and considers the advantages of using transformer models. A method for detecting intrusions in computer networks is proposed. This method differs from known approaches by utilizing the Vision Transformer for Small-size Datasets (ViTSD) deep learning algorithm. The method incorporates procedures to reduce the correlation of input data and transform data into a specific format required for model operations. The developed methods are implemented using Python and the GOOGLE COLAB cloud service with Jupyter Notebook. Conclusions. Experiments confirmed the efficiency of the proposed method. The use of the developed method based on the ViTSD algorithm and the data preprocessing procedure increases the model's accuracy to 98.7\%. This makes it possible to recommend it for practical use, in order to improve the accuracy of identifying the state of a computer system.}, language = {en} } @article{MarcosZambranoLopezMolinaBakirGungoretal.2023, author = {Marcos-Zambrano, Laura Judith and L{\´o}pez-Molina, V{\´i}ctor Manuel and Bakir-Gungor, Burcu and Frohme, Marcus and Karaduzovic-Hadziabdic, Kanita and Klammsteiner, Thomas and Ibrahimi, Eliana and Lahti, Leo and Loncar-Turukalo, Tatjana and Dhamo, Xhilda and Simeon, Andrea and Nechyporenko, Alina and Pio, Gianvito and Przymus, Piotr and Sampri, Alexia and Trajkovik, Vladimir and Lacruz-Pleguezuelos, Blanca and Aasmets, Oliver and Araujo, Ricardo and Anagnostopoulos, Ioannis and Aydemir, {\"O}nder and Berland, Magali and Calle, M. Luz and Ceci, Michelangelo and Duman, Hatice and G{\"u}ndoğdu, Aycan and Havulinna, Aki S. and Kaka Bra, Kardokh Hama Najib and Kalluci, Eglantina and Karav, Sercan and Lode, Daniel and Lopes, Marta B. and May, Patrick and Nap, Bram and Nedyalkova, Miroslava and Paci{\^e}ncia, In{\^e}s and Pasic, Lejla and Pujolassos, Meritxell and Shigdel, Rajesh and Sus{\´i}n, Antonio and Thiele, Ines and Truică, Ciprian-Octavian and Wilmes, Paul and Yilmaz, Ercument and Yousef, Malik and Claesson, Marcus Joakim and Truu, Jaak and Carrillo de Santa Pau, Enrique}, title = {A toolbox of machine learning software to support microbiome analysis}, series = {Frontiers in Microbiology}, volume = {14}, journal = {Frontiers in Microbiology}, publisher = {Frontiers}, issn = {1664-302X}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-18271}, year = {2023}, abstract = {The human microbiome has become an area of intense research due to its potential impact on human health. However, the analysis and interpretation of this data have proven to be challenging due to its complexity and high dimensionality. Machine learning (ML) algorithms can process vast amounts of data to uncover informative patterns and relationships within the data, even with limited prior knowledge. Therefore, there has been a rapid growth in the development of software specifically designed for the analysis and interpretation of microbiome data using ML techniques. These software incorporate a wide range of ML algorithms for clustering, classification, regression, or feature selection, to identify microbial patterns and relationships within the data and generate predictive models. This rapid development with a constant need for new developments and integration of new features require efforts into compile, catalog and classify these tools to create infrastructures and services with easy, transparent, and trustable standards. Here we review the state-of-the-art for ML tools applied in human microbiome studies, performed as part of the COST Action ML4Microbiome activities. This scoping review focuses on ML based software and framework resources currently available for the analysis of microbiome data in humans. The aim is to support microbiologists and biomedical scientists to go deeper into specialized resources that integrate ML techniques and facilitate future benchmarking to create standards for the analysis of microbiome data. The software resources are organized based on the type of analysis they were developed for and the ML techniques they implement. A description of each software with examples of usage is provided including comments about pitfalls and lacks in the usage of software based on ML methods in relation to microbiome data that need to be considered by developers and users. This review represents an extensive compilation to date, offering valuable insights and guidance for researchers interested in leveraging ML approaches for microbiome analysis.}, language = {en} } @article{D'EliaTruuLahtietal.2023, author = {D'Elia, Domenica and Truu, Jaak and Lahti, Leo and Berland, Magali and Papoutsoglou, Georgios and Ceci, Michelangelo and Zomer, Aldert and Lopes, Marta B. and Ibrahimi, Eliana and Gruca, Aleksandra and Nechyporenko, Alina and Frohme, Marcus and Klammsteiner, Thomas and Carrillo de Santa Pau, Enrique and Marcos-Zambrano, Laura Judith and Hron, Karel and Pio, Gianvito and Simeon, Andrea and Suharoschi, Ramona and Moreno-Indias, Isabel and Temko, Andriy and Nedyalkova, Miroslava and Apostol, Elena-Simona and Truică, Ciprian-Octavian and Shigdel, Rajesh and Telalović, Jasminka Hasić and Bongcam-Rudloff, Erik and Przymus, Piotr and Jordamović, Naida Babić and Falquet, Laurent and Tarazona, Sonia and Sampri, Alexia and Isola, Gaetano and P{\´e}rez-Serrano, David and Trajkovik, Vladimir and Klucar, Lubos and Loncar-Turukalo, Tatjana and Havulinna, Aki S. and Jansen, Christian and Bertelsen, Randi J. and Claesson, Marcus Joakim}, title = {Advancing microbiome research with machine learning: key findings from the ML4Microbiome COST action}, series = {Frontiers in Microbiology}, volume = {14}, journal = {Frontiers in Microbiology}, publisher = {Frontiers}, issn = {1664-302X}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-18004}, year = {2023}, abstract = {The rapid development of machine learning (ML) techniques has opened up the data-dense field of microbiome research for novel therapeutic, diagnostic, and prognostic applications targeting a wide range of disorders, which could substantially improve healthcare practices in the era of precision medicine. However, several challenges must be addressed to exploit the benefits of ML in this field fully. In particular, there is a need to establish "gold standard" protocols for conducting ML analysis experiments and improve interactions between microbiome researchers and ML experts. The Machine Learning Techniques in Human Microbiome Studies (ML4Microbiome) COST Action CA18131 is a European network established in 2019 to promote collaboration between discovery-oriented microbiome researchers and data-driven ML experts to optimize and standardize ML approaches for microbiome analysis. This perspective paper presents the key achievements of ML4Microbiome, which include identifying predictive and discriminatory 'omics' features, improving repeatability and comparability, developing automation procedures, and defining priority areas for the novel development of ML methods targeting the microbiome. The insights gained from ML4Microbiome will help to maximize the potential of ML in microbiome research and pave the way for new and improved healthcare practices.}, language = {en} } @article{SchillingBeyerleinChien2023, author = {Schilling, Vincent and Beyerlein, Peter and Chien, Jeremy}, title = {A Bioinformatics Analysis of Ovarian Cancer Data Using Machine Learning}, series = {Algorithms}, volume = {16}, journal = {Algorithms}, number = {7}, publisher = {MDPI}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-17751}, year = {2023}, abstract = {The identification of biomarkers is crucial for cancer diagnosis, understanding the underlying biological mechanisms, and developing targeted therapies. In this study, we propose a machine learning approach to predict ovarian cancer patients' outcomes and platinum resistance status using publicly available gene expression data. Six classical machine-learning algorithms are compared on their predictive performance. Those with the highest score are analyzed by their feature importance using the SHAP algorithm. We were able to select multiple genes that correlated with the outcome and platinum resistance status of the patients and validated those using Kaplan-Meier plots. In comparison to similar approaches, the performance of the models was higher, and different genes using feature importance analysis were identified. The most promising identified genes that could be used as biomarkers are TMEFF2, ACSM3, SLC4A1, and ALDH4A1.}, language = {en} } @article{ErmakovaFabianGolimblevskaiaetal.2023, author = {Ermakova, Tatiana and Fabian, Benjamin and Golimblevskaia, Elena and Henke, Max}, title = {A Comparison of Commercial Sentiment Analysis Services}, series = {SN Computer Science}, volume = {4}, journal = {SN Computer Science}, publisher = {Springer Nature}, issn = {2661-8907}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-17598}, year = {2023}, abstract = {Empirical insights into promising commercial sentiment analysis solutions that go beyond the claims of their vendors are rare. Moreover, due to the constant evolution in the field, previous studies are far from reflecting the current situation. The goal of this article is to evaluate and compare current solutions using two experimental studies. In the first part of the study, based on tweets about airline service quality, we test the solutions of six vendors with different market power, such as Amazon, Google, IBM, Microsoft, Lexalytics, and MeaningCloud, and report their measures of accuracy, precision, recall, (macro)F1, time performance, and service level agreements (SLA). Furthermore, we compare two of the services in depth with multiple data sets and over time. The services tested here are Google Cloud Natural Language API and MeaningCloud Sentiment Analysis API. For evaluating the results over time, we use the same data set as in November 2020. In addition, further topic-specific and general Twitter data sets are used. The experiments show that the IBM Watson NLU and Google Cloud Natural Language API solutions may be preferred when negative text detection is the primary concern. When tested in July 2022, the Google Cloud Natural Language API was still the clear winner compared to the MeaningCloud Sentiment Analysis API, but only on the airline service quality data set; on the other data sets, both services provided specific benefits and drawbacks. Furthermore, we detected changes in the sentiment classification over time with both services. Our results motivate that an independent, critical, and longitudinal experimental analysis of sentiment analysis services can provide interesting insights into their overall reliability and particular classification accuracy beyond marketing claims to critically compare solutions based on real data and analyze potential weaknesses and margins of error before making an investment.}, language = {en} } @article{SiptrothMoskalenkoKrumbiegeletal.2023, author = {Siptroth, Julienne and Moskalenko, Olga and Krumbiegel, Carsten and Ackermann, J{\"o}rg and Koch, Ina and Pospisil, Heike}, title = {Investigation of metabolic pathways from gut microbiome analyses regarding type 2 diabetes mellitus using artificial neural networks}, series = {Discover Artificial Intelligence}, volume = {3}, journal = {Discover Artificial Intelligence}, number = {1}, publisher = {Springer Nature}, issn = {2731-0809}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-17497}, year = {2023}, abstract = {Background Type 2 diabetes mellitus is a prevalent disease that contributes to the development of various health issues, including kidney failure and strokes. As a result, it poses a significant challenge to the worldwide healthcare system. Research into the gut microbiome has enabled the identification and description of various diseases, with bacterial pathways playing a critical role in this context. These pathways link individual bacteria based on their biological functions. This study deals with the classification of microbiome pathway profiles of type 2 diabetes mellitus patients. Methods Pathway profiles were determined by next-generation sequencing of 16S rDNA from stool samples, which were subsequently assigned to bacteria. Then, the involved pathways were assigned by the identified gene families. The classification of type 2 diabetes mellitus is enabled by a constructed neural network. Furthermore, a feature importance analysis was performed via a game theoretic approach (SHapley Additive exPlanations). The study not only focuses on the classification using neural networks, but also on identifying crucial bacterial pathways. Results It could be shown that a neural network classification of type 2 diabetes mellitus and a healthy comparison group is possible with an excellent prediction accuracy. It was possible to create a ranking to identify the pathways that have a high impact on the model prediction accuracy. In this way, new associations between the alteration of, e.g. a biosynthetic pathway and the presence of diabetes mellitus type 2 disease can also be discovered. The basis is formed by 946 microbiome pathway profiles from diabetes mellitus type 2 patients (272) and healthy comparison persons (674). Conclusion With this study of the gut microbiome, we present an approach using a neural network to obtain a classification of healthy and type 2 diabetes mellitus and to identify the critical features. Intestinal bacteria pathway profiles form the basis.}, language = {en} } @article{LangenbergMaErmakovaetal.2023, author = {Langenberg, Anna and Ma, Shih-Chi and Ermakova, Tatiana and Fabian, Benjamin}, title = {Formal Group Fairness and Accuracy in Automated Decision Making}, series = {Mathematics}, volume = {11}, journal = {Mathematics}, number = {8}, publisher = {MDPI}, issn = {2227-7390}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-17323}, year = {2023}, abstract = {Most research on fairness in Machine Learning assumes the relationship between fairness and accuracy to be a trade-off, with an increase in fairness leading to an unavoidable loss of accuracy. In this study, several approaches for fair Machine Learning are studied to experimentally analyze the relationship between accuracy and group fairness. The results indicated that group fairness and accuracy may even benefit each other, which emphasizes the importance of selecting appropriate measures for performance evaluation. This work provides a foundation for further studies on the adequate objectives of Machine Learning in the context of fair automated decision making.}, language = {en} } @article{AlekseevaNechyporenkoFrohmeetal.2023, author = {Alekseeva, Victoriia and Nechyporenko, Alina and Frohme, Marcus and Gargin, Vitaliy and Meniailov, Ievgen and Chumachenko, Dmytro}, title = {Intelligent Decision Support System for Differential Diagnosis of Chronic Odontogenic Rhinosinusitis Based on U-Net Segmentation}, series = {Electronics}, volume = {12}, journal = {Electronics}, number = {5}, publisher = {MDPI}, issn = {2079-9292}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-17088}, year = {2023}, abstract = {The share of chronic odontogenic rhinosinusitis is 40\% among all chronic rhinosinusitis. Using automated information systems for differential diagnosis will improve the efficiency of decision-making by doctors in diagnosing chronic odontogenic rhinosinusitis. Therefore, this study aimed to develop an intelligent decision support system for the differential diagnosis of chronic odontogenic rhinosinusitis based on computer vision methods. A dataset was collected and processed, including 162 MSCT images. A deep learning model for image segmentation was developed. A 23 convolutional layer U-Net network architecture has been used for the segmentation of multi-spiral computed tomography (MSCT) data with odontogenic maxillary sinusitis. The proposed model is implemented in such a way that each pair of repeated 3 × 3 convolutions layers is followed by an Exponential Linear Unit instead of a Rectified Linear Unit as an activation function. The model showed an accuracy of 90.09\%. To develop a decision support system, an intelligent chatbot allows the user to conduct an automated patient survey and collect patient examination data from several doctors of various profiles. The intelligent information system proposed in this study made it possible to combine an image processing model with a patient interview and examination data, improving physician decision-making efficiency in the differential diagnosis of Chronic Odontogenic Rhinosinusitis. The proposed solution is the first comprehensive solution in this area.}, language = {en} } @article{KoenigKiffnerKuhlsetal.2023, author = {K{\"o}nig, Hannes J. and Kiffner, Christian and Kuhls, Katrin and Uthes, Sandra and Harms, Verena and Wieland, Ralf}, title = {Planning for wolf-livestock coexistence: landscape context predicts livestock depredation risk in agricultural landscapes}, series = {animal}, volume = {17}, journal = {animal}, number = {3}, publisher = {Elsevier}, issn = {1751-7311}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-17032}, year = {2023}, abstract = {Extensive pastoral livestock systems in Central Europe provide multiple ecosystem services and support biodiversity in agricultural landscapes but their viability is challenged by livestock depredation (LD) associated with the recovery of wolf populations. Variation in the spatial distribution of LD depends on a suite of factors, most of which are unavailable at the appropriate scales. To assess if LD patterns can be predicted sufficiently with land use data alone at the scale of one federal state in Germany, we employed a machine-learning-supported resource selection approach. The model used LD monitoring data, and publicly available land use data to describe the landscape configuration at LD and control sites (resolution 4 km * 4 km). We used SHapley Additive exPlanations to assess the importance and effects of landscape configuration and cross-validation to evaluate the model performance. Our model predicted the spatial distribution of LD events with a mean accuracy of 74\%. The most influential land use features included grassland, farmland and forest. The risk of livestock depredation was high if these three landscape features co-occurred with a specific proportion. A high share of grassland, combined with a moderate proportion of forest and farmland, increased LD risk. We then used the model to predict the LD risk in five regions; the resulting risk maps showed high congruence with observed LD events. While of correlative nature and lacking specific information on wolf and livestock distribution and husbandry practices, our pragmatic modelling approach can guide spatial prioritisation of damage prevention or mitigation practices to improve livestock-wolf coexistence in agricultural landscapes.}, language = {en} } @misc{Behrens2022, type = {Master Thesis}, author = {Behrens, Alena}, title = {Evaluation des Sprachmodells GPT-3 f{\"u}r den Einsatz an der ZBW - Leibniz Informationszentrum Wirtschaft}, doi = {10.15771/MA_2022_4}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-16940}, school = {Technische Hochschule Wildau}, pages = {156}, year = {2022}, abstract = {Die Verarbeitung nat{\"u}rlicher Sprache durch Maschinen ist fester Bestandteil des Alltags. Die Verarbeitung erfolgt durch technische Sprachmodelle, die bereits viele Aufgaben erfolgreich l{\"o}sen. Durch die stetige Weiterentwicklung entstehen m{\"a}chtigere Modelle mit neuen F{\"a}higkeiten. Eines der neuesten ist GPT-3. Das Ziel dieser Masterarbeit ist es, das Sprachmodell GPT-3 und dessen praktischen Einsatz zu untersuchen. Zur Untersuchung des Modells werden die Fragen gestellt, was GPT-3 von anderen Modellen unterscheidet, wie es funktioniert und in welchen Bereichen es eingesetzt werden kann. Weiterhin wird die Frage untersucht, ob das Modell f{\"u}r den geplanten Chatbot der ZBW - Leibniz Informationszentrum Wirtschaft genutzt werden kann. Zudem wird untersucht, welche Voraussetzungen f{\"u}r einen Einsatz des Modells in einem Chatbot n{\"o}tig sind. F{\"u}r die ersten drei Forschungsfragen wurde eine Literaturrecherche angewendet. Damit wurden Hintergr{\"u}nde zum Modell zusammengetragen, sowie Berichte aus der Praxis ber{\"u}cksichtigt. F{\"u}r den Einsatz bei einem Chatbot wurde ein Prototyp erstellt, der von den Testpersonen genutzt und bewertet wurde. Die Auswertung aktueller Literatur zu GPT-3 zeigt, dass sich das Modell durch seine neue Funktionsweise, die Gr{\"o}ße der Parameter und Testdaten von anderen abhebt. Dadurch wurden neue F{\"a}higkeiten m{\"o}glich. GPT-3 kann bei vielen Einsatzgebieten, die auf nat{\"u}rlicher Sprache beruhen, eingesetzt werden. Dies umfasst klassische Natural Language Processing-Aufgaben, sowie neue Bereiche wie etwa die Programmierung. Im praktischen Test zeigte sich, dass es eine hohe Pr{\"a}zision erfordert, das Modell auf neue Themen anzulernen. Die Quote der richtigen Antwort des Chatbots war zu gering, um den Einsatz uneingeschr{\"a}nkt empfehlen zu k{\"o}nnen. Es bedarf weiterer Untersuchungen, um das Modell f{\"u}r den Einsatz im Chatbot der ZBW zu optimieren.}, language = {de} } @article{ChumachenkoButkevychLodeetal.2022, author = {Chumachenko, Dmytro and Butkevych, Mykola and Lode, Daniel and Frohme, Marcus and Schmailzl, Kurt J. G. and Nechyporenko, Alina}, title = {Machine Learning Methods in Predicting Patients with Suspected Myocardial Infarction Based on Short-Time HRV Data}, series = {Sensors}, volume = {22}, journal = {Sensors}, number = {18}, publisher = {MDPI}, issn = {1424-8220}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-16521}, year = {2022}, abstract = {Diagnosis of cardiovascular diseases is an urgent task because they are the main cause of death for 32\% of the world's population. Particularly relevant are automated diagnostics using machine learning methods in the digitalization of healthcare and introduction of personalized medicine in healthcare institutions, including at the individual level when designing smart houses. Therefore, this study aims to analyze short 10-s electrocardiogram measurements taken from 12 leads. In addition, the task is to classify patients with suspected myocardial infarction using machine learning methods. We have developed four models based on the k-nearest neighbor classifier, radial basis function, decision tree, and random forest to do this. An analysis of time parameters showed that the most significant parameters for diagnosing myocardial infraction are SDNN, BPM, and IBI. An experimental investigation was conducted on the data of the open PTB-XL dataset for patients with suspected myocardial infarction. The results showed that, according to the parameters of the short ECG, it is possible to classify patients with a suspected myocardial infraction as sick and healthy with high accuracy. The optimized Random Forest model showed the best performance with an accuracy of 99.63\%, and a root mean absolute error is less than 0.004. The proposed novel approach can be used for patients who do not have other indicators of heart attacks.}, language = {en} } @article{WielandKuhlsLentzetal.2021, author = {Wieland, Ralf and Kuhls, Katrin and Lentz, Hartmut H.K. and Conraths, Franz and Kampen, Helge and Werner, Doreen}, title = {Combined climate and regional mosquito habitat model based on machine learning}, series = {Ecological Modelling}, volume = {452}, journal = {Ecological Modelling}, address = {Elsevier}, issn = {0304-3800}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-15649}, year = {2021}, abstract = {Besides invasive mosquito species also several native species are proven or suspected vectors of arboviruses as West Nile or Usutu virus in Western Europe. Habitat models of these native vectors can be a helpful tool for assessing the risk of autochthonous occurrence, outbreaks and spread of diseases caused by such arboviruses. Modelling native mosquitoes is complicated because of the perfect adaptation to the climatic and landscape conditions and their high abundance in contrast to invasive species. Here we present a new approach for such a habitat model for native mosquito species in Germany, which are considered as vectors of West Nile virus (WNV). Epizootic emergence of WNV was registered in Germany since 2018. The models are based on surveillance data of mosquitoes from the German citizen science project "M{\"u}ckenatlas" complemented by data from systematic trap monitoring in Germany, and on data freely available from the Deutscher Wetterdienst (DWD) and OpenStreetMap (OSM). While climatic factors still play an important role, we could show that habitat suitability is predictable only by the combination of the climate model with a regional model. Both models were based on a machine-learning approach using XGBoost. Evaluation of the accuracy of the models was done by statistical analysis, determining among others feature importances using the SHAP-Library. Final output of the combined climatic and regional models are maps showing the superposed habitat suitability which are generated through a number of steps described in detail. These maps also include the registered cases of WNV infections in the selected region of Germany.}, language = {en} } @article{RosendorffHodesFabian2021, author = {Rosendorff, Andr{\´e} and Hodes, Alexander and Fabian, Benjamin}, title = {Artificial intelligence for last-mile logistics - Procedures and architecture}, series = {The Online Journal of Applied Knowledge Management (OJAKM)}, volume = {9}, journal = {The Online Journal of Applied Knowledge Management (OJAKM)}, number = {1}, publisher = {The International Institute for Applied Knowledge Management}, issn = {2325-4688}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-15586}, pages = {46 -- 61}, year = {2021}, abstract = {Artificial Intelligence (AI) is becoming increasingly important in many industries due to its diverse areas of application and potential. In logistics in particular, increasing customer demands and the growth in shipment volumes are leading to difficulties in forecasting delivery times, especially for the last mile. This paper explores the potential of using AI to improve delivery forecasting. For this purpose, a structured theoretical solution approach and a method for improving delivery forecasting using AI are presented. In doing so, the important phases of the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework, a standard process for data mining, are adopted and discussed in detail to illustrate the complexity and importance of each task such as data preparation or evaluation. Subsequently, by embedding the described solution into an overall system architecture for information systems, ideas for the integration of the solution into the complexity of real information systems for logistics are given.}, language = {en} } @inproceedings{ErmakovaHenkeFabian2021, author = {Ermakova, Tatiana and Henke, Max and Fabian, Benjamin}, title = {Commercial Sentiment Analysis Solutions: A Comparative Study}, publisher = {SciTePress}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-15509}, pages = {103 -- 114}, year = {2021}, abstract = {Empirical insights into high-promising commercial sentiment analysis solutions that go beyond their vendors' claims are rare. Moreover, due to ongoing advances in the field, earlier studies are far from reflecting the current situation due to the constant evolution of the field. The present research aims to evaluate and compare current solutions. Based on tweets on the airline service quality, we test the solutions of six vendors with different market power, such as Amazon, Google, IBM, Microsoft, and Lexalytics, and MeaningCloud, and report their measures of accuracy, precision, recall, (macro) F1, time performance, and service level agreements (SLA). For positive and neutral classifications, none of the solutions showed precision of over 70\%. For negative classifications, all of them demonstrate high precision of around 90\%, however, only IBM Watson NLU and Google Cloud Natural Language achieve recall of over 70\% and thus can be seen as worth considering for application scenarios w here negative text detection is a major concern. Overall, our study shows that an independent, critical experimental analysis of sentiment analysis services can provide interesting insights into their general reliability and particular classification accuracy beyond marketing claims to critically compare solutions based on real-world data and analyze potential weaknesses and margins of error before making an investment.}, language = {en} } @inproceedings{vandeSandCorasanitiReiffStephan2021, author = {van de Sand, Ron and Corasaniti, Sandra and Reiff-Stephan, J{\"o}rg}, title = {Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers}, series = {2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021)}, volume = {1}, booktitle = {2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021)}, publisher = {TIB Open Publishing}, address = {Hannover}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:526-opus4-14965}, pages = {155 -- 164}, year = {2021}, abstract = {Chiller systems are used in many different applications in both the industrial and the commercial sector. They are considered major energy consumers and thus contribute a non-negligible factor to environmental pollution as well as to the overall operating cost. In addition, chillers, especially in industrial applications, are often associated with high reliability requirements, as unplanned system downtimes are usually costly. As many studies over the past decades have shown, the presence of faults can lead to significant performance degradation and thus higher energy consumption of these systems. Thus, data-driven fault detection plays an ever-increasing role in terms of energy efficient control strategies. However, labelled data to train associated algorithms are often only available to a limited extent, which consequently inhibits the broad application of such technologies. Therefore, this paper presents an approach that exploits only a small amount of labelled and large amounts of unlabelled data in the training phase in order to detect fault related anomalies. For this, the model utilizes the residual space of the data transformed through principal component analyses in conjunction with a biased support vector machine, which can be ascribed to the concept of semi-supervised learning, or more specifically, positive-unlabelled learning.}, language = {en} }