Refine
Document Type
- Master's Thesis (7) (remove)
Has Fulltext
- yes (7)
Is part of the Bibliography
- no (7)
Keywords
- Maschinelles Lernen (7) (remove)
Institute
Open Access
- Closed Access (4)
- Closed (3)
As e-commerce platforms have grown in popularity, new difficulties have emerged, such as the growing use of bots—automated programs—to engage with e-commerce websites. Even though some algorithms are helpful, others are malicious and can seriously hurt e-commerce platforms by making fictitious purchases, posting fictitious evaluations, and gaining control of user accounts. Therefore, the development of more effective and precise bot identification systems is urgently needed to stop such actions. This thesis proposes a methodology for detecting bots in E-commerce using machine learning algorithms such as K-nearest neighbors, Decision Tree, Random Forest, Support Vector Machine, and Neural Network. The purpose of the research is to assess and contrast the output of these machine learning methods. The suggested approach will be based on data that is readily accessible to the public, and the study’s focus will be on the research of bots in e-commerce.
The purpose of the study is to provide an overview of bots in e-commerce, as well as information on the different kinds and traits of bots, as well as current research on bots in e-commerce and associated work on bot detection in e-commerce. The research also seeks to create a more precise and effective bot detection system as well as find critical factors in detecting bots in e-commerce.
This research is significant because it sheds light on the increasing issue of bots in e-commerce and the requirement for more effective bot detection systems. The suggested approach for using machine learning algorithms to identify bots in ecommerce can give e-commerce platforms a more precise and effective bot detection system to stop malicious bot activities. The study’s results can also be used to create a more effective bot detection system and pinpoint key elements in detecting bots in e-commerce.
Das Verstehen und Extrahieren von Informationen aus Dokumenten stellt eine Herausforderung dar, welche den Einsatz weiterer Technologien bedarf. Vorliegende
Masterarbeit untersucht die Anwendbarkeit von Methoden des maschinellen Lernens im Bereich der Wissensextraktion auf Basis von Angebotsdokumenten. Hierbei gilt die Frage zu klären, inwiefern sich diese Dokumente eignen, um Strukturen
für die Modellierung mit einem Produktkonfigurator zu lernen. Kern der Arbeit stellen die Datenaufbereitung von PDF-Dokumenten sowie das Modeling multimodal
lernender Algorithmen dar. Abgesehen von Texten werden zusätzlich Layoutinformationen für das Lernen der Strukturen genutzt. Zudem werden die Ergebnisse der
erstellten Modelle evaluiert und die Güte in Anbetracht des vorliegenden Problems
bewertet.
Mit der prototypischen Implementierung einer automatisierten Dokumentengenerierung wird demonstriert, wie das extrahierte Wissen in der Software CAS Configurator Merlin genutzt werden kann.
The research employed HPTLC Pro System and other HPTLC instruments from CAMAG® to conduct various laboratory tests, aiming to compile a database for subsequent analyses. Utilizing MATLAB, distinct codes were developed to reveal patterns within analyzed biomasses and pyrolysis oils (sewage sludge, fermentation residue, paper sludge, and wood). Through meticulous visual and numerical analysis, shared characteristics among different biomasses and their respective pyrolysis oils were revealed, showcasing close similarities within each category. Notably, minimal disparity was observed in fermentation residue and wood biomasses with a similarity coefficient of 0.22. Similarly, for pyrolysis oils, the minimal disparity was found in fermentation residues 1 and 3, with a disparity coefficient of 1.41. Despite higher disparity coefficients in certain results, specific biomasses and pyrolysis oils, such as fermentation residue and sewage sludge, exhibited close similarities, with disparity coefficients of 0.18 and 0.55, respectively. The database, derived from triplicate experimentation, now serves as a valuable resource for rapid analysis of newly acquired raw materials. Additionally, the utility of HPTLC PRO as an investigation tool, enabling simultaneous analysis of up to five samples, was emphasized, although areas for improvement in derivatization methods were identified.
Organizations striving to achieve success in the long term must have a positive brand image which will have direct implications on the business. In the face of the rising cyber threats and intense competition, maintaining a threat-free domain is an important aspect of preserving that image in today's internet world. Domain names are often near-synonyms for brand names for numerous companies. There are likely thousands of domains that try to impersonate the big companies in a bid to trap unsuspecting users, usually falling prey to attacks such as phishing or watering hole. Because domain names are important for organizations for running their business online, they are also particularly vulnerable to misuse by malicious actors. So, how can you ensure that your domain name is protected while still protecting your brand identity? Brand Monitoring, for example, may assist. The term "Brand Monitoring" applies only to keep tabs on an organization's brand performance, reception, and overall online presence through various online channels and platforms [1]. There has been a rise in the need of maintaining one's domain clear of any linkages to malicious activities as the threat environment has expanded. Since attackers are targeting domain names of organizations and luring unsuspecting users to visit malicious websites, domain monitoring becomes an important aspect. Another important aspect of brand abuse is how attackers leverage brand logos in creating fake and phishing web pages. In this Master Thesis, we try to solve the problem of classification of impersonated domains using rule-based and machine learning algorithms and automation of domain monitoring. We first use a rule-based classifier and Machine Learning algorithms to classify the domains gathered into two buckets – "Parked" and "Non-Parked". In the project's second phase, we will deploy object detection models (Scale Invariant Feature Transform - SIFT and Multi-Template Matching – MTM) to detect brand logos from the domains of interest.
In der vorliegenden Thesis werden Empfehlungsalgorithmen zur Verbesserung von Wein-Empfehlungen evaluiert. Der Algorithmus wird zur Entscheidung zwischen zwei Weinen eingesetzt, so dass der jeweils für den Kunden geeignetere Wein empfohlen wird. Das derzeitige System setzt Collaborative Filtering durch den Alternating Least Squares (ALS) Algorithmus um. Bei Kunden und Weinen, die nicht die notwendigen Bedingungen für die Anwendung von ALS erfüllen, wird durch Zufall entschieden.
Dem bestehenden Ansatz wurden folgenden Verfahren gegenübergestellt: Content-based Filtering mit einen Autoencoder und Hybrid Filtering mit einem neuronalen Netz sowie mit der Empfehlungsbibliothek LightFM. Die neuen Ansätze berücksichtigen immer die Weineigenschaften und können für noch nicht gekaufte Weine eingesetzt werden (Cold-Start Problem). Verglichen wurden die Ansätze durch zwei Ranking-Methoden und einen selbst-entwickelten offline A/B-Test.
Unter den neuen Ansätzen schnitt LightFM am besten ab. ALS lieferte insgesamt die besten Ranking-Werte. Durch ein online A/B-Test zwischen ALS und LightFM konnten keine signifikanten Ergebnisse ermittelt werden. Insgesamt konnte auf Basis der in den Tests gesammelten Daten keine Verbesserung der Empfehlungslogik gegenüber dem bestehenden Verfahren mittels ALS nachgewiesen werden. Für eine abschließende statistisch signifikante Beurteilung müssten mehr online A/B-Tests durchgeführt werden.
Much of the research in the field of audio-based machine learning has focused on recreating human speech via feature extraction and imitation, known as deepfakes. The current state of affairs has prompted a look into other areas, such as the recognition of recording devices, and potentially speakers, by only analysing sound files. Segregation and feature extraction are at the core of this approach.
This research focuses on determining whether a recorded sound can reveal the recording device with which it was captured. Each specific microphone manufacturer and model, among other characteristics and imperfections, can have subtle but compounding effects on the results, whether it be differences in noise, or the recording tempo and sensitivity of the microphone while recording. By studying these slight perturbations, it was found to be possible to distinguish between microphones based on the sounds they recorded.
After the recording, pre-processing, and feature extraction phases we completed, the prepared data was fed into several different machine learning algorithms, with results ranging from 70% to 100% accuracy, showing Multi-Layer Perceptron and Logistic Regression to be the most effective for this type of task.
This was further extended to be able to tell the difference between two microphones of the same make and model. Achieving the identification of identical models of a microphone suggests that the small deviations in their manufacturing process are enough of a factor to uniquely distinguish them and potentially target individuals using them. This however does not take into account any form of compression applied to the sound files, as that may alter or degrade some or most of the distinguishing features that are necessary for this experiment.
Building on top of prior research in the area, such as by Das et al. in in which different acoustic features were explored and assessed on their ability to be used to uniquely fingerprint smartphones, more concrete results along with the methodology by which they were achieved are published in this project’s publicly accessible code repository.
The progress in machine learning has led to advanced deep neural networks. These networks are widely used in computer vision tasks and safety-critical applications. The automotive industry, in particular, has experienced a significant transformation with the integration of deep learning techniques and neural networks. This integration contributes to the realization of autonomous driving systems. Object detection is a crucial element in autonomous driving. It contributes to vehicular safety and operational efficiency. This technology allows vehicles to perceive and identify their surroundings. It detects objects like pedestrians, vehicles, road signs, and obstacles. Object detection has evolved from being a conceptual necessity to an integral part of advanced driver assistance systems (ADAS) and the foundation of autonomous driving technologies. These advancements enable vehicles to make real-time decisions based on their understanding of the environment, improving safety and driving experiences. However, the increasing reliance on deep neural networks for object detection and autonomous driving has brought attention to potential vulnerabilities within these systems. Recent research has highlighted the susceptibility of these systems to adversarial attacks. Adversarial attacks are well-designed inputs that exploit weaknesses in the deep learning models underlying object detection. Successful attacks can cause misclassifications and critical errors, posing a significant threat to the functionality and safety of autonomous vehicles. With the rapid development of object detection systems, the vulnerability to adversarial attacks has become a major concern. These attacks manipulate inputs to deceive the target system, significantly compromising the reliability and safety of autonomous vehicles. In this study, we focus on analyzing adversarial attacks on state-of-the-art object detection models. We create adversarial examples to test the models’ robustness. We also check if the attacks work on a different object detection model meant for similar tasks. Additionally, we extensively evaluate recent defense mechanisms to see how effective they are in protecting deep neural networks (DNNs) from adversarial attacks and provide a comprehensive overview of the most commonly used defense strategies against adversarial attacks, highlighting how they can be implemented practically in real-world situations.