Estimate Train Driving Time with Artifical Inteligence
Erik Kostelansky, Emil Kršák, Tomas Kello · 2019
The main goal of this paper is to analyze properties of train set due to driving time with the use of machine learning. The implementation of the regression machine learning model is done using the framework ML.NET in the programming language C#. The beginning of the thesis is dedicated to a theoretical overview about machine learning and an introduction to the framework ML.NET. Then we will use the aforementioned framework to achieve the best results while predicting the driving time of trains. We need to analyze the data set before attempting to implement the model. The process of detect missing data and making a strategy about is also done before implementing the model. This process is followed by finding the list of train features, which greatly affect the final predicted driving time of the train. In conclusion we evaluate the acquired results and the pros and cons of implementing a machine learning model in the programming language C# using the framework ML.NET.