Refine
Year of publication
Document Type
- Master's Thesis (5)
- Article (2)
- Bachelor Thesis (1)
- Study Thesis (1)
- Working Paper (1)
Language
- English (10) (remove)
Has Fulltext
- yes (10)
Keywords
- Business Intelligence (2)
- Visualisierung (2)
- API (1)
- Bestärkendes Lernen <Künstliche Intelligenz> (1)
- Big Data (1)
- Client-Server-Konzept (1)
- Clinical Trials (1)
- Data Mining (1)
- Data Ware House (1)
- Data-Warehouse-Konzept (1)
Faculty
- Fakultät 10 / Institut für Informatik (10) (remove)
This study paper introduces different tools, i.e. analytical methods and visualizations, in business intelligence environments. It especially emphasizes the use of OLAP-based technologies as a tradtional kind of data analysis in contrast to as graph analysis and formal concept analysis as rather new approaches in the area of visual analytics.
The amount of data produced and stored in multiple types of distributed data sources is growing steadily. A crucial factor that determines whether data can be analyzed efficiently is the use of adequate visualizations. Almost simultaneously with the ongoing availability of data numerous types of visualization techniques have emerged. Since ordinary business intelligence users typically lack expert visualization knowledge, the selection and creation of visualizations can be a very time- and knowledge-consuming task. To encounter these problems an architecture that aims at supporting ordinary BI users in the selection of adequate visualizations is developed in this thesis. The basic idea is to automatically provide visualization recommendations based on the concrete BI scenario and formalized visualization knowledge. Ontologies that formalize all relevant knowledge play an important role in the developed architecture and are the key to make the knowledge machine-processable.
The topic for the thesis originated from the CAP4ACCESS project run by the European Commission and its partners, which deals towards the sensiti-zation of people and development of tools for awareness about people with movement disabilities. The explorative analysis is never ending and to explore and find interest-ing patterns and the results is a tedious task. Therefore, a scientific approach was very important. To start with, familiarizing the domain and the data sources were done. Thereafter, selection of methodology for data analysis was done which resulted in the use of CRISP-DM methodology. The data sources are the source of blood to the analysis methodology, and as there were two sources of data that is MICROM and OSM Wheelchair History(OWH), it was important to integrate them together to extract relevant datasets. Therefore a functional and technically impure data warehouse was created, from which the datasets are extracted and analysed.The next task was to select appropriate tools for analysis. This task was very important as the data set although was not big data but con-tained a large number of rows. After careful analysis, Apache spark and its machine learning library were utilized for building and testing supervised models. DataFrame API for Python, Pandas, the machine learning library Sci-kit learn provided unsupervised algorithms for analysis, the association rule analysis was performed using WEKA. Tableau[21] and Matplotlib[24] provide attractive visualizations for representation and analysis.
An empirical evaluation of using the Swift language as the underlying technology of RESTful APIs
(2016)
The purpose of the current thesis is to determine the appropriateness of using the Swift language as the underlying technology for the development of RESTful APIs in a Linux environment. The current paper describes the process of designing, implementing and testing individual RESTful API components based on Node.js, PHP, Python and Swift and seeks to determine whether Swift is a viable alternative.
The thesis begins by defining a methodology for implementing and testing individual RESTful API components based on Node.js, PHP, Python and Swift. It then proceeds to detail the implementation and testing processes, following with an analytic discussion regarding the advantages and drawbacks of using the Swift language as the underlying technology for RESTful APIs and server-side Linux-based applications in general.
Based on the implementation process and on the results of the previously mentioned evaluation phase, it can be stated that the Swift language is not yet ready to be used in a production environment. However, its rapid evolution and potential for surpassing its competitors in the foreseeable future make it an ideal candidate for implementing RESTful APIs to be used in development environments.
This thesis is aimed for finding a solution for non-gaming application of Virtual Reality technology in data visualization and analysis. Starting by reconstructing the concept of Virtual Reality, the paper then describes the principles, concepts and techniques of designing a Virtual Reality application. In the last part of the thesis, a detailed description of how a prototype implemented is presented to provide a preview of how data visualization and analysis and Virtual Reality technology can be combined together in order to enable users to perceive and comprehend data in a possibly better way.
Observational studies and clinical trials have become increasingly important over recent years and play an essential role in advancing medical knowledge. In today’s world of clinical research, it is not possible to imagine trials without the founda-tion of a well-established it-infrastructure. Electronic capture and usage of data is pervasive.
In practice, medical progress requires the ability to integrate data from different systems. An essential factor in enabling different actors, such as institutions and hospitals, to have their systems exchange structured data and make use of the information is the interoperability of the data and systems.
FHIR (Fast Healthcare Interoperable Resources) is a free and easily customizable HL7 platform standard, based on 30 years of experience of HL7. It is focused on providing health-related information and defines a set of capabilities used in the health care process.
This thesis will provide a conceptual approach for working with FHIR, as well as concrete approaches for working with FHIR profiles and for customizing the standard for particular use cases. It will be carried out in cooperation with the Medical Systems R&D, which is a service provider within the University Hospital of Cologne.
The guiding request approach will focus on the evaluation of requirements for clini-cal trials and how clinical research protocols can be represented in an interoperable and machine-parsable format using FHIR.
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.
Academic search systems aid users in finding information covering specific topics of scientific interest and have evolved from early catalog-based library systems to modern web-scale systems. However, evaluating the performance of the underlying retrieval approaches remains a challenge. An increasing amount of requirements for producing accurate retrieval results have to be considered, e.g., close integration of the system’s users. Due to these requirements, small to mid-size academic search systems cannot evaluate their retrieval system in-house. Evaluation infrastructures for shared tasks alleviate this situation. They allow researchers to experiment with retrieval approaches in specific search and recommendation scenarios without building their own infrastructure. In this paper, we elaborate on the benefits and shortcomings of four state-of-the-art evaluation infrastructures on search and recommendation tasks concerning the following requirements: support for online and offline evaluations, domain specificity of shared tasks, and reproducibility of experiments and results. In addition, we introduce an evaluation infrastructure concept design aiming at reducing the shortcomings in shared tasks for search and recommender systems.
The demand for explainable and transparent models increases with the continued success of reinforcement learning. In this article, we explore the potential of generating shallow decision trees (DTs) as simple and transparent surrogate models for opaque deep reinforcement learning (DRL) agents. We investigate three algorithms for generating training data for axis-parallel and oblique DTs with the help of DRL agents (“oracles”) and evaluate these methods on classic control problems from OpenAI Gym. The results show that one of our newly developed algorithms, the iterative training, outperforms traditional sampling algorithms, resulting in well-performing DTs that often even surpass the oracle from which they were trained. Even higher dimensional problems can be solved with surprisingly shallow DTs. We discuss the advantages and disadvantages of different sampling methods and insights into the decision-making process made possible by the transparent nature of DTs. Our work contributes to the development of not only powerful but also explainable RL agents and highlights the potential of DTs as a simple and effective alternative to complex DRL models.