Clustering Variations within Ferry Operational Patterns by applying Unsupervised Machine Learning - A Case Study

This paper investigates the process of applying unsupervised machine learning on operational data from ferry M/F Langeland. Previous unsupervised machine learning studies have, inter alia, focused on energy emission patterns in buildings through the application of transfer learning; encoding of time series to images; convolutional autoencoders; and clustering algorithms. This study follows a very similar approach into the field of ferry operational patterns, with the primary focus of identifying differences within maneuvering patterns. The findings unequivocally show that it is possible to identify such patterns, albeit inconsistencies exist which are almost certainly due to external factors.