C Mathematical and Quantitative Methods
Refine
Document Type
- Bachelor Thesis (2)
- Master's Thesis (2)
- Study Thesis (1)
Has Fulltext
- yes (5)
Keywords
- API (1)
- Client-Server-Konzept (1)
- Datenbank (1)
- Datenverarbeitung (1)
- ERP (1)
- HTTP (1)
- Hypermedia (1)
- Informatik (1)
- Künstliche Intelligenz (1)
- Leistungsbewertung (1)
Faculty
- Fakultät 10 / Institut für Informatik (5) (remove)
As a customer, it can be frustrating to face an empty shelf in a store. The market does not always realize that a product has been out of stock for a while, as the item is still listed as in stock in the inventory management system. To address this issue, a camera should be used to check for Out-of-Stock (OOS) situations.
This master thesis evaluates different model configurations of Artificial Neural Networks (ANNs) to determine which one best detects OOS situations in the market using images. To create a dataset, 2,712 photos were taken in six stores. The photos clearly show whether there is a gap on the shelf or if the product is in stock. Based on the pre-trained VGG16 model from Keras, two fully connected layers were implemented, with 36 different ANNs differing in the optimization method and activation function pairings. In total, 216 models were generated in this thesis to investigate the effects of three different optimization methods combined with twelve different activation function pairings. An almost balanced ratio of OOS and in-stock data was used to generate these models.
The evaluation of the generated OOS models shows that the FTRL optimization method achieved the least favorable results and is therefore not suitable for this application. Model configurations using the Adam or SGD optimization methods achieve much better results. Of the top six model configurations, five use the Adam optimization method and one uses SGD. They all achieved an accuracy of at least 93% and were able to predict the Recall for the OOS class with at least 91%.
As the data ratio between OOS and in-stock data did not correspond to reality in the previously generated models, the in-stock images were augmented. Including the augmented images, new OOS models were generated for the top six model configurations. The results of these OOS models show no convergences. This suggests that more epochs in the training phase lead to better results. However, the results of the OOS model using the Adam optimization method and the Sigmoid and ReLU activation functions stand out positively. It achieved the best result with an accuracy of 97.91% and a Recall of the OOS class of 87.82%.
Overall, several OOS models have the potential to increase both market sales and customer satisfaction. In a future study, the OOS models should be installed in the market to evaluate their performance under real conditions. The resulting insights can be used for continuous optimization of the model.
Das Ziel der vorliegenden Bachelorarbeit war es, Technologietrends von Enterprise-Resource-Planning (ERP) Systemen zu definieren und anhand eines modernen Produktes vorzustellen. Ein ERP-System unterstützt Unternehmen in der Organisation und Durchführung von sämtlichen Geschäftsprozessen. Zur Untersuchung der wünschenswerten Eigenschaften eines ERP-Systems wurden anhand von Literaturarbeit Anforderungen definiert. Dabei wurde deutlich, dass die Verarbeitung von sämtlichen Geschäftsdaten in Echtzeit einen hohen Stellenwert hat. Hierfür bietet das ERP-System S/4HANA mit verschieden Technologien eine Lösung auf dem Datenbanksystem an. So zeigt ein direkter Performancevergleich zwischen den Datenbanksystemen von S/4HANA und einem aktuellen Konkurrenzprodukt, dass das Verarbeiten von großen Datenmengen in kürzester Zeit möglich ist. Diese Arbeit ist an Fachkreise der Wirtschaftsinformatik und Personengruppen mit Interesse an ERP und Datenbanken Systemen gerichtet.
REST became the go to approach when it comes to large scale distributed systems on, or outside the World Wide Web. This paper aims to give a brief overview of what REST is and what its main draws and benefits are. Secondly, I will showcase the implementation of REST using HTTP and why this approach became as popular as it is today. Based on my research I concluded that REST’s advantages in scalability, coupling, performance and its seamless integration with HTTP enabled it to rightfully overtake classic RPC based approaches.
Vergleich verschiedener Lernmethoden neuronaler Netze bei der Analyse von „Social Media“ Inhalten
(2018)
The goal of this bachelor thesis was the comparison of different learning methods in neural networks. The methods were applied to detect hate posts on social media plat-forms like twitter. To achieve this, a supervised Recurrent Neural Network and a self-supervised Word2Vec model were implemented. The results of both implementations show the importance of choosing the correct dataset and a learning method generating significant results. The problems of both implementations were identified and formulated into possible solutions to achieve more accurate predictions in future. This thesis is of high interest for students and developers in the area of sentiment analysis.
Das Ziel der vorliegenden Masterthesis ist es, einen Überblick der verschiedenen Datenbanktypen und Leistungsanalysen zu geben. Die vergleichende Literaturstudie beschäftigt sich mit einem jungen Forschungsfeld und betrachtet insbesondere nichtrelationale NoSQL-Datenbanken, welche in den letzten Jahren immer beliebter geworden sind und einige Vorteile gegenüber relationalen Datenbanken aufweisen. Doch was können die konkreten Datenbankimplementierungen bei unterschiedliche Datenmodellen leisten und welcher Testaufbau bietet sich bei welchen Einsatzanforderungen an? Zu Anfang definiert diese Arbeit Kriterien zur Bewertung von Leistung und untersucht experimentelle Vorgehensweise verschiedener Forscher. Ein wichtiger Fokus liegt darauf, die Vergleichbarkeit der Messmethoden und Ergebnisse einzuschätzen und zu gewährleisten. Neben dem methodischen Vorgehen wird mit dem YCSB-Framework ein wichtiges Werkzeug besprochen, mit dem Leistungsmessungen in NoSQL-Datenbanken implementiert werden können.