Building a Machine Learning Environment

OData support
Supervisor:
Dr. Forstner Bertalan
Department of Automation and Applied Informatics

In recent years an increasing number of industry equipment and machines are equipped with a multitude of sensors. With the data provided by these sensors, businesses have an opportunity to optimize their processes, increase their efficiency and solve other critical challenges.

Due to the vast amount and the process-time sensitivity of data incoming from the sensors the only solution is the use of machine learning algorithms. For the use of these algorithms a software architecture is necessary which satisfies the industry standard requirements (high availability, scalability, efficiency, cost effectivity) and is capable of storing the data, running the machine learning models and providing the output of the models for other systems.

In this paper I propose an architecture designed for the above requirements, I present a machine learning model which can be integrated into the architecture and I examine already existing systems in the railway industry, and their impact on the industry.

Downloads

Please sign in to download the files of this thesis.